Reappraisal of Austrosaurus Mckillopi Longman, 1933 from the Allaru

Home , Aeolosaurini, Agustinia, Austrosaurus, Cerro Barcino Formation, Daxiatitan, Diamantinasaurus

1 Reappraisal of Austrosaurus mckillopi Longman, 1933 from the

2 Allaru Mudstone of Queensland, Australia’s first named

3 Cretaceous sauropod dinosaur

5 STEPHEN F. POROPAT, JAY P. NAIR, CAITLIN E. SYME, PHILIP D. MANNION,

6 PAUL UPCHURCH, SCOTT A. HOCKNULL, ALEX G. COOK, TRAVIS R. TISCHLER

7 and TIMOTHY HOLLAND

9 POROPAT, S. F., NAIR, J. P., SYME, C. E., MANNION, P. D., UPCHURCH, P., HOCKNULL, S. A.,

10 COOK, A. G., TISCHLER, T.R. & HOLLAND, T. 201#. Reappraisal of Austrosaurus mckillopi

11 Longman, 1933 from the Allaru Mudstone of Queensland, Australia’s first named Cretaceous

12 sauropod dinosaur. Alcheringa XX, XX–XX. ISSN ###.

14 Austrosaurus mckillopi Longman, 1933 was the first Cretaceous sauropod reported from

15 Australia, and the first Cretaceous dinosaur reported from Queensland (northeast Australia).

16 This sauropod taxon was established on the basis of several fragmentary presacral vertebrae

17 (QM F2316) derived from the uppermost Lower Cretaceous (upper Albian) Allaru Mudstone,

18 at a locality situated 77 km west-northwest of Richmond, Queensland. Prior to its rediscovery

19 in 2014, the type site was considered lost after failed attempts to relocate it in the 1970s.

20 Excavations at the site in 2014 and 2015 led to the recovery of several partial dorsal ribs and

21 fragments of presacral vertebrae, all of which clearly pertained to a single sauropod dinosaur.

22 The discovery of new material of the type individual of Austrosaurus mckillopi, in tandem

23 with a reassessment of the material collected in the 1930s, has facilitated the rearticulation of

24 the specimen. The resultant vertebral series comprises six presacral vertebrae—the

25 posteriormost cervical and five anteriormost dorsals—in association with five left dorsal ribs

26 and one right one. The fragmentary nature of the type specimen has historically hindered

27 assessments of the phylogenetic affinities of Austrosaurus, as has the fact that these

28 evaluations were often based on a subset of the type material. The reappraisal of the type

29 series of Austrosaurus presented herein, on the basis of both external morphology and

30 internal morphology visualised through CT data, validates it as a diagnostic titanosauriform

31 taxon, tentatively placed in Somphospondyli, and characterised by the possession of an

32 accessory lateral pneumatic foramen on dorsal vertebra I (a feature which appears to be

33 autapomorphic) and by the presence of a robust ventral midline ridge on the centra of dorsal

34 vertebrae I and II. The interpretation of the anteriormost preserved vertebra in Austrosaurus

35 as a posterior cervical has also prompted the re-evaluation of an isolated, partial, posterior

36 cervical vertebra (QM F6142, the “Hughenden sauropod”) from the upper Albian Toolebuc

37 Formation (which underlies the Allaru Mudstone). Although this vertebra preserves an

38 apparent unique character of its own (a spinopostzygapophyseal lamina fossa), it is not able

39 to be referred unequivocally to Austrosaurus and is retained as Titanosauriformes indet.

40 Austrosaurus mckillopi is one of the oldest known sauropods from the Australian Cretaceous

41 based on skeletal remains, and potentially provides phylogenetic and/or palaeobiogeographic

42 context for later taxa such as Wintonotitan wattsi, Diamantinasaurus matildae and

43 Savannasaurus elliottorum.

45 Stephen F. Poropat [[email protected]; [email protected]] (corresponding

46 author), Department of Chemistry and Biotechnology, Swinburne University of Technology,

47 John St, Hawthorn, Victoria 3122, Australia and Australian Age of Dinosaurs Museum of

48 Natural History, Lot 1 Dinosaur Drive, PO Box 408, Winton, Queensland 4735, Australia;

49 Jay P. Nair [[email protected]; [email protected]] School of Biological Sciences, The

50 University of Queensland, St Lucia, Queensland 4072, Australia; Caitlin E. Syme

51 [[email protected]] School of Biological Sciences, The University of

52 Queensland, St Lucia, Queensland 4072, Australia; Philip D. Mannion

53 [[email protected]] Department of Earth Science and Engineering, Imperial

54 College London, South Kensington Campus, London SW7 2AZ, United Kingdom; Paul

55 Upchurch [[email protected]] Department of Earth Sciences, University College

56 London, Gower Street, London WC1E 6BT, United Kingdom; Scott A. Hocknull

57 [[email protected]] Geosciences, Queensland Museum, 122 Gerler Rd, Hendra,

58 Queensland 4011, Australia; Alex G. Cook [[email protected]] School of Earth

59 Sciences, The University of Queensland, St Lucia, Queensland 4072, Australia; Travis R.

60 Tischler [[email protected]] Australian Age of Dinosaurs Museum of Natural

61 History, Lot 1 Dinosaur Drive, PO Box 408, Winton, Queensland 4735, Australia; Timothy

62 Holland [[email protected]] Kronosaurus Korner, 91 Goldring St, Richmond,

63 Queensland 4822, Australia. Received XX.X.201X; revised XX.X.201X; accepted XX.X.201X.

65 Key words: Austrosaurus; Dinosauria; Sauropoda; Titanosauriformes; Australia; Cretaceous;

66 Gondwana.

68 SAUROPOD dinosaur fossils were not reported from Australia until 1926, despite the

69 fact that the first discovery of an Australian sauropod specimen was made in 1913. This

70 incomplete humerus (QM F311)—found in Cretaceous strata near the town of Blackall in

71 Queensland—was not determined to pertain to a dinosaur, let alone a sauropod, until 1980

72 (Molnar 2001b). The first Australian sauropod specimen that was recognised as such was

73 discovered in 1924 near the town of Roma, Queensland. This partial skeleton (QM F1659)

74 was designated the type of Rhoetosaurus brownei and remains one of only two Jurassic

75 sauropod specimens known from Australia (Longman 1926, 1927a, b, 1929, Thulborn 1985,

76 Rich & Vickers-Rich 2003, Nair & Salisbury 2012). The other, a distal caudal vertebra

77 (UWA 82468), was found in the mid-1970s in Middle Jurassic rocks near Geraldton, Western

78 Australia (Long 1992).

79 Cretaceous sauropod remains are only known from three Australian states. Footprints

80 found in the Broome Sandstone (Valanginian–Barremian) of the Dampier Peninsula

81 constitute the only evidence of Cretaceous sauropods in Western Australia (Thulborn et al.

82 1994, Thulborn 2012, Salisbury et al. 2017), whereas opalised sauropod teeth (AM F66769,

83 AM F66770) from the Griman Creek Formation (middle Albian) of Lightning Ridge are all

84 that has been reported from New South Wales (Molnar & Salisbury 2005, Molnar 2011b). In

85 Queensland, the Griman Creek Formation has yielded two possible sauropod elements: a

86 fragmentary right ischium (QM F54817) and a possible vertebral fragment (QM F11043;

87 Molnar 2011b).

88 The Eromanga Basin in Queensland has produced the overwhelming majority of

89 Australia’s sauropod skeletal remains. Rare specimens have been described from the

90 Toolebuc Formation and the Allaru Mudstone, both of which constitute upper Albian marine

91 deposits (Molnar & Salisbury 2005). However, the bulk of the known sauropod remains from

92 the Eromanga Basin derive from the youngest Cretaceous unit preserved therein, the

93 uppermost Albian–lower Turonian Winton Formation (Coombs & Molnar 1981, Molnar &

94 Salisbury 2005, Greentree 2011, Bryan et al. 2012, Tucker et al. 2013, Tucker 2014), which

95 evinces the existence of a vast floodplain. To date, three sauropod taxa have been reported

96 from the Winton Formation: the non-titanosaurian somphospondylan Wintonotitan wattsi

97 (Hocknull et al. 2009, Poropat et al. 2015a); and the titanosaurs Diamantinasaurus matildae

98 (Hocknull et al. 2009, Poropat et al. 2015b, 2016) and Savannasaurus elliottorum (Poropat et

99 al. 2016). However, the first Cretaceous sauropod, and indeed the first Cretaceous dinosaur,

100 reported from Queensland was not from the Winton Formation: it was found in the Allaru

101 Mudstone, and bears the name Austrosaurus mckillopi (Longman 1933).

102 ===PLEASE INSERT FIGURE 1===

103 The type and only known specimen of Austrosaurus mckillopi was found in August

104 1932 on Clutha sheep station, 77 km west-northwest of Richmond, Queensland (Fig. 1).

105 Clutha overseer Henry Burgoyne Wade (Fig. 2A) found fragments of fossilised bone near the

106 southwest corner of Whitewood Paddock. He showed these to the station manager, Harley

107 John McKillop (Fig. 2B), who contacted his brother, Dr Martin Joseph McKillop (Fig. 2C).

108 Shortly afterwards, M. J. McKillop travelled from Brisbane to Clutha, helped H. B. Wade and

109 H. J. McKillop to recover additional specimens, and sent a sketch of one to Heber Albert

110 Longman (Fig. 2D), then director of the Queensland Museum. Longman requested that the

111 specimens be sent to Brisbane, and they arrived in January 1933. By March 1933, Longman

112 had correctly determined that the bones derived from a sauropod dinosaur and made them the

113 type specimen of Austrosaurus mckillopi, honouring M. J. McKillop with the species epithet

114 (Longman 1933). A sign was erected by H. B. Wade and H. J. McKillop at the type site to

115 demarcate its significance (Fig. 2E).

116 ===PLEASE INSERT FIGURE 2===

117 In his original description of Austrosaurus mckillopi, Longman only mentioned three

118 specimens (and figured two), all of which were presacral vertebrae (catalogued as QM

119 F2316). It is probable that Longman received more than three specimens in the initial

120 shipment from Clutha, given that contemporary newspaper articles stated that six vertebrae

121 were preserved (Anonymous 1933b, a). However, only on the three specimens described by

122 Longman is there evidence of “repeated soakings in shellac solution” (Longman 1933, p.

123 132). Correspondence between H. J. McKillop and Longman confirms that a second

124 shipment of specimens was dispatched from Clutha to the Queensland Museum in June 1933.

125 Molnar (2001b, p. 141, 2010, p. 423, 2011a, p. 332) and Molnar & Salisbury (2005, p. 456)

126 suggested that this second shipment comprised five large and more than ten small pieces.

127 However, in his letter to Longman on 23/05/1933, H. J. McKillop stated that the additional

128 specimens were “smaller fossils from the same animal”, which suggests that all eight large

129 portions of Austrosaurus vertebrae had been delivered to the Queensland Museum in the first

130 shipment: furthermore, this implies that Longman did not describe all of the specimens at his

131 disposal in March 1933. The total number of blocks presently catalogued under QM F2316 is

132 25: eight large, and 17 small (Appendix 1).

133 Longman never visited Clutha sheep station, possibly because H. B. Wade and H. J.

134 McKillop left the property when it was sold in July 1933. Evidently, only two attempts were

135 made by palaeontologists to revisit the Austrosaurus type locality prior to 2014. Tony

136 Thulborn (then at The University of Queensland) and Mary Wade (then at the Queensland

137 Museum) attempted to relocate the site in 1976 (as alluded to by Molnar 1982b, p. 622, 2010,

138 p. 423), immediately prior to their first excavation at the Lark Quarry Dinosaur Stampede

139 (Thulborn & Wade 1979, 1984); however, they failed to find additional specimens (R. A.

140 Thulborn, pers. comm. 2015). Ralph Molnar (then at the University of New South Wales), his

141 wife Barbara, and Peter Bell (then a graduate student) visited Clutha in late June 1977, but

142 they also failed to find any additional specimens (R. E. Molnar, pers. comm. 2016).

143 Perfunctory assessments of the vertebrae of Austrosaurus collected in the 1930s

144 suggested that they represented part of an articulated series. This strongly implied that more

145 of the same skeleton was preserved but remained unexcavated (as also suggested by Molnar

146 2010). Despite this, aside from a few brief considerations (Molnar 2001b, Molnar &

147 Salisbury 2005, Molnar 2010, 2011a), the bulk of the Austrosaurus material has remained

148 undescribed, and the specimen has never been rearticulated.

149 Longman (1933) published a map of Clutha Station (provided to him by H. J.

150 McKillop) which included the paddock fence lines and an “X” marking the Austrosaurus site

151 in the southwest corner of Whitewood Paddock (Fig. 3A). This map was overlaid onto

152 Google Earth satellite images, which revealed that the fences on Clutha had not been moved

153 since the 1930s. This implied that a search for the site could be constrained to a small section

154 of one paddock. Clutha sheep station straddles two of the geological maps produced by the

155 Bureau of Mineral Resources (Vine et al. 1963, 1970), both of which concur that the non-

156 Quaternary sedimentary rocks exposed on the property, including the Austrosaurus type site,

157 fall entirely within the bounds of the Allaru Mudstone (Fig. 3B).

158 ===PLEASE INSERT FIGURE 3===

159 In December 1933, M. J. McKillop sent Longman a photograph of a sign, supported

160 by two “gidgee” (Acacia) posts, which was erected by H. B. Wade and H. J. McKillop at the

161 Austrosaurus site shortly after the taxon was described (Fig. 2E). No trace of the sign has

162 been found; it is presumed to have disintegrated. However, the posts did not disintegrate, and

163 they were known to John Wharton (mayor of Richmond Shire), who grew up on Clutha.

164 Wharton and the last author (TH) attempted to relocate the site at ground level in early 2014.

165 Although this proved futile, Wharton succeeded in finding the posts from the air with a

166 helicopter. Nearby, he found fragments of mudstone which contained fossilised bone with

167 camellate internal texture, showing nearly identical preservation to the type specimen of

168 Austrosaurus mckillopi.

169 In July 2014, a team led by TH and SFP visited the presumed Austrosaurus site and

170 recovered additional fragments of sauropod bone (rib and camellate internal vertebral

171 fragments) from the topsoil. This was followed by a small-scale excavation in August 2014,

172 wherein three fragmentary ribs were found next to each other. One rib portion was left in situ

173 and covered with plaster in anticipation of a subsequent excavation. This transpired in July

174 2015, and the presence of six ribs was confirmed—three of these connected to rib portions

175 excavated in 2014. All specimens were collected and donated to the Kronosaurus Korner

176 Marine Fossil Museum (Richmond, Queensland), where they are presently on display.

177 The left ribs were preserved with their medial surfaces facing upwards and their

178 tapered anterior margins directed to the northwest. A single right dorsal rib was found on top

179 of the posteriormost left dorsal ribs preserved. The spacing between the left ribs was

180 consistent with them having been derived from an articulated skeleton (as indicated by the

181 vertebral centra originally described by Longman); however, no additional sauropod remains

182 were found, despite extensive excavation around, and below the level of, the ribs. The

183 immediate site is now considered to be exhausted of fossilised skeletal material, with the type

184 of Austrosaurus having been augmented as much as possible. However, it is not infeasible

185 that additional portions of the carcass might eventually be found further distant from the

186 already excavated area. A comprehensive account of the discovery and rediscovery of the

187 Austrosaurus site has been published elsewhere (Poropat 2016).

188 In this paper the augmented type specimen of Austrosaurus mckillopi is reappraised,

189 and the first full description of this taxon is provided based on both external and internal

190 characteristics (the latter having been visualised from CT data). The taphonomic processes

191 which affected the Austrosaurus type specimen post mortem are considered and the

192 implications of its preservation in a marine setting are explored. Comparisons of the type

193 specimen of Austrosaurus with other sauropods from both Australia and elsewhere support

194 the notion that it is a distinct, diagnostic taxon. This in turn has facilitated a reassessment of

195 its phylogenetic placement, and has raised questions over its palaeobiogeographic and

196 phylogenetic significance for other, geologically younger Australian sauropod taxa and for

197 Early Cretaceous South American titanosauriforms. Finally, the possible referral to

198 Austrosaurus mckillopi of the enigmatic “Hughenden sauropod”, represented by a single,

199 incomplete, posterior cervical vertebra (QM F6142) from the Toolebuc Formation, is

200 assessed.

201

202 Institutional abbreviations: AM, Australian Museum, Sydney, New South Wales, Australia;

203 AOD, Australian Age of Dinosaurs Natural History Museum, Winton, Queensland, Australia;

204 KK, Kronosaurus Korner Marine Fossil Museum, Richmond, Queensland, Australia;

205 MMCH, Museo Municipal ‘Ernesto Bachman’, Villa El Chocón, Neuquén, Argentina; MN,

206 Museu Nacional, Rio de Janeiro, Brazil; MUCPv, Museo de Geología y Paleontología de la

207 Universidad Nacional del Comahue, Neuquén, Argentina; QM, Queensland Museum,

208 Brisbane, Queensland, Australia; UWA, University of Western Australia, Perth, Western

209 Australia.

210

211 Geological and depositional setting and associated palaeofauna

212 QM F2316 was preserved in the Allaru Mudstone, the second youngest of the marine

213 units preserved within the Eromanga Basin (Fig. 1). Much of the Allaru Mudstone was

214 deposited below wave base in a low-energy marine setting, and comprises blue–grey

215 mudstones and siltstones, some of which are calcareous (Exon & Senior 1976). However, its

216 lowermost and uppermost sections are sporadically coarser-grained and contain sedimentary

217 structures such as planar cross-bedding, hummocky cross-stratification and ripple cross-

218 lamination, which suggests both initial and terminal shallowing during deposition. In the

219 northern Eromanga Basin the Allaru Mudstone conformably overlies the Toolebuc

220 Formation, whereas in the southern part of the basin (where the Toolebuc Formation is

221 absent) it lies directly upon the upper Aptian–middle Albian Wallumbilla Formation (Gray et

222 al. 2002). The Allaru Mudstone is conformably overlain by the marine Mackunda Formation,

223 which is in turn overlain by the paralic–terrestrial Winton Formation.

224 Sauropod remains from the Allaru Mudstone are rare, with few specimens other than

225 Austrosaurus mckillopi known (Poropat et al. 2014). The only other terrestrial tetrapod

226 remains reported from the unit are: three ornithopod specimens (Molnar 1980, 1982b, 1984b,

227 Lees 1986, Molnar 1996a, Agnolin et al. 2010), one of which has been designated

228 Muttaburrasaurus sp. (Molnar 1996a); three ankylosaur specimens (Molnar 1984b, 1996b,

229 Leahey & Salisbury 2013, Leahey et al. 2015), including the holotype specimen of

230 Kunbarrasaurus ieversi [Leahey et al. 2015; formerly Minmi sp. (Molnar 1996b, Molnar &

231 Clifford 2000, Molnar 2001a, Molnar & Clifford 2001)]; and fragmentary cranial elements

232 which have been attributed to a late-surviving dicynodont (Longman 1916, Thulborn &

233 Turner 2003) but also apparently compare favourably with baurusuchian crocodylomorphs

234 (Agnolin et al. 2010, p. 293).

235 Unsurprisingly, remains of ancient marine vertebrates are far more common in the

236 Allaru Mudstone than are those of terrestrial vertebrates. Polycotylid (Mobbs 1990, Hughes

237 2003), elasmosaurid (Persson 1960, Kear 2003, McHenry et al. 2005) and pliosaurid

238 plesiosaurs (Holland 2015) are all represented, alongside ichthyosaurs (Longman 1943, Wade

239 1984, 1990, Zammit et al. 2010, Kear & Zammit 2014) and marine turtles (SFP and TH, pers.

240 obs.). A variety of chondrichthyan and actinopterygian fishes has also been recovered from

241 the Allaru Mudstone (Longman 1913, 1932, Bardack 1962, Bartholomai 1969, Lees &

242 Bartholomai 1987, Kemp 1991, Bartholomai 2004, 2008, 2010b, 2012, Wretman & Kear

243 2014).

244 The Allaru Mudstone hosts a diverse fossil mollusc fauna, including squids (Wade

245 1993), ammonites (Day 1969, McNamara 1978, Henderson & Kennedy 2002, Henderson &

246 McKenzie 2002), belemnites (Cook 2012), scaphopods (Stilwell 1999) and bivalves (Cook

247 2012). Non-molluscan invertebrates, such as crustaceans [brachyurans (Etheridge 1892,

248 Woodward 1892, Etheridge 1917, Woods 1953, Glaessner 1980) and ostracods

249 (Krömmelbein 1975)] and echinoderms (Cook 2008, 2012) are locally abundant, whereas

250 corals are rare (Jell et al. 2011). Foraminifera from the Allaru Mudstone (Playford et al.

251 1975) indicate that the water temperature was cool, that the water salinity was lower than that

252 of normal seawater, and that the seaway was shallow, probably less than 100 m deep (Haig

253 1979a, b). The abundance of planktonic organisms is also indicative of near-normal marine

254 conditions (Exon & Senior 1976), whereas the abundance and diversity of benthic organisms

255 (Haig 1980, 1982) appears to indicate well-oxygenated waters, at least during the early

256 phases of deposition (Haig & Lynch 1993). Calcareous nannofossils suggest that the

257 deposition of the lowermost Allaru Mudstone (at least) took place at a relatively high

258 palaeolatitude (~55°; Seton et al. 2012) and/or in an environment characterised by cool near-

259 surface water temperatures (Shafik 1985), an interpretation supported by analyses of isotope

260 ratios in belemnite rostra (~19°C; Price et al. 2012).

261 Although terrestrial plant fossils are rare in the Allaru Mudstone [the cycadale

262 Nilssonia mucronatum is one of the few described specimens (Rozefelds 1986)], studies of

263 the palynomorphs have shown that this unit lies within the upper Coptospora paradoxa and

264 Phimopollenites pannosus palynological zones (Burger 1986). The Allaru Mudstone is,

265 consequently, ascribed a late Albian age, which is further reinforced by the presence of the

266 ammonite Goodhallites goodhallites (Henderson & Kennedy 2002).

267

268 Systematic Palaeontology

269 DINOSAURIA Owen, 1842

270 SAURISCHIA Seeley, 1887

271 SAUROPODOMORPHA von Huene, 1932

272 SAUROPODA Marsh, 1878

273 EUSAUROPODA Upchurch, 1995

274 NEOSAUROPODA Bonaparte, 1986

275 MACRONARIA Wilson and Sereno, 1998

276 TITANOSAURIFORMES Salgado et al., 1997

277 ?SOMPHOSPONDYLI Wilson and Sereno, 1998

278

279 Austrosaurus mckillopi Longman, 1933

280

281 Holotype. QM F2316: four partial, articulated presacral vertebrae, preserved within three

282 blocks (Longman 1933: “specimens A–C”), comprising the posterior portion of the centrum

283 of the posteriormost cervical vertebra, the centra and partial neural arches of dorsal vertebrae

284 I and II, and a fragment of the centrum of dorsal vertebra III.

285 Hypodigm. QM F2316 and KK F1020, comprising the holotype and additional specimens

286 referrable to the type individual: a series of articulated presacral vertebrae, comprising the

287 posteriormost cervical vertebra and dorsal vertebrae I–V, associated with six dorsal ribs (left

288 ribs I–V, right rib IV) and numerous additional fragments. QM F2316 comprises the six

289 vertebrae and multiple associated fragments (including one small rib portion; Appendix 1),

290 whereas KK F1020 comprises the six dorsal ribs, as well as additional rib and vertebra

291 fragments.

292 Type locality. Southwest corner of Whitewood Paddock, Clutha Station (Fig. 3), ~55 km

293 north-northwest of Maxwelton (~77 km west-northwest of Richmond), north-central

294 Queensland, Australia (Fig. 1).

295 Type horizon. Allaru Mudstone (Lower Cretaceous; upper Albian).

296 Original diagnosis. “Dorsal vertebrae markedly opisthocoelous; centra with thin cortical

297 walls, much enlarged at the enarthrodial articulations; intramural region a complex of small

298 cavities; pleurocoeles[sic] prominent, with external and internal divisions. Neural arch with

299 deep recess between the prezygapophyseal lamina and the infradiapophysial buttress”

300 (Longman 1933, p. 132).

301 Comments on original diagnosis. Opisthocoelous anterior dorsal vertebrae are now

302 recognised as being characteristic of most eusauropods (Wilson & Sereno 1998, Wilson

303 2002), whereas opisthocoelous middle–posterior dorsal vertebrae, although typical among

304 macronarians (Wilson 2002), are also present in some non-neosauropod sauropods

305 (Carballido et al. 2011b, p. 634). The combination of thin exterior walls, small internal

306 cavities and prominent lateral pneumatic foramina (= pleurocoels) in the anterior dorsal

307 vertebrae exemplifies the camellate internal texture now widely recognised in the presacral

308 vertebrae of titanosauriform sauropods (Wedel 2003). Finally, several sauropod taxa are now

309 known to possess a deep fossa [centrodiapophyseal fossa sensu Wilson et al. (2011b)] like

310 that seen on Longman’s “specimen A” between the “prezygapophyseal lamina” [= anterior

311 centroparapophyseal lamina (ACPL)] and the “infradiapophysial buttress” [= posterior

312 centrodiapophyseal lamina (PCDL)]. In sum, none of Longman’s (1933) characters can be

313 considered to be either autapomorphic or differentially diagnostic for Austrosaurus because

314 all are more widely distributed within Sauropoda. However, these characters do suggest that

315 Austrosaurus is a titanosauriform sauropod.

316 Revised diagnosis. A titanosauriform sauropod characterised as follows [potential

317 autapomorphies are indicated with an asterisk (*)]: dorsal vertebrae I–II with ventral ridges

318 flanked by shallow, circular fossae; dorsal vertebra I with accessory lateral pneumatic

319 foramen situated anterodorsal to the parapophysis*; dorsal rib distal ends in cross-section

320 ranging from plank-like (I–III) to semi-plank-like (IV) to subcircular (V).

321

322 Description

323 ===PLEASE INSERT FIGURE 4===

324 Prior to our work on Austrosaurus, the number of vertebrae comprising the type series

325 had never been established with certainty, nor had the vertebral column ever been

326 successfully rearticulated. The type specimen of Austrosaurus was stated by Coombs &

327 Molnar (1981, p. 358) to comprise “a series of six fragmentary dorsal vertebrae”, although

328 further elaboration was not made. Molnar & Salisbury (2005, p. 456) and Molnar (2010, p.

329 423) claimed that at least eight vertebrae were present, although no evidence was presented to

330 support this beyond the fact that there were eight large fragments containing portions of

331 presacral vertebrae catalogued under QM F2316. Prior to the rearticulation of the specimen,

332 all that could be stated with certainty was that five condyle-cotyle pairs were catalogued as

333 part of the type series of Austrosaurus, indicating the presence of a minimum of six

334 vertebrae.

335 The rearticulation of the vertebral column, successfully undertaken by SFP and JPN,

336 revealed that the three specimens described by Longman (1933) were found to articulate with

337 one another, with the undescribed condyle–cotyle pairs forming a sequence succeeding those

338 three (Appendix 1). The preserved presacral series, as articulated, comprises the

339 posteriormost cervical vertebra, and dorsal vertebrae I–V (Appendix 1; Fig. 4–8). On the left

340 lateral side of the preserved vertebral column (Fig. 5), portions of two fragmentary ribs are

341 preserved, not far removed from their in vivo positions. The recovery of several associated,

342 effectively in situ left dorsal ribs from the Austrosaurus type site in 2014–2015 (Fig. 9)

343 suggests that the carcass was buried with its left side lying on the seafloor (all of the left ribs

344 were preserved with their medial surfaces up; Fig. 10). All but one of the right ribs are

345 missing; it is likely that any others which were preserved were lost to erosion long before the

346 specimen was discovered in 1932.

347 A fragmentary ammonite (registered as QM F2321) was found in association with

348 Austrosaurus, whereas other mollusc specimens [Inoceramus (Bivalvia: Cryptodonta) and

349 Beudanticeras (Cephalopoda: Ammonoidea)] remain embedded within matrix adhered to

350 QM F2316 (Longman 1933, Molnar 2010). At least three additional ammonites, not visible

351 on any exposed surface, have been identified through observation of the CT data. This

352 associated invertebrate fauna clearly demonstrates that Austrosaurus was preserved in a

353 marine setting.

354 Presacral vertebrae: general patterns. All preserved presacral vertebrae of Austrosaurus

355 mckillopi are strongly opisthocoelous. The high degree of weathering to which the majority

356 of the specimens has been subjected has exposed, and often accentuated, the camellate

357 internal texture of the vertebrae. All of the vertebrae of Austrosaurus bear deeply-penetrating

358 pneumatic foramina set within fossae (Fig. 5, 6). Few portions of the neural arches are

359 preserved.

360 Several morphological changes are evident along the vertebral column. The first two

361 dorsal vertebrae each possess a well-developed ventral median ridge, bounded on either side

362 by circular depressions (Fig. 8); these ridges and depressions are absent in the succeeding

363 dorsal vertebrae. The size of the vertebral condyle decreases in each successive vertebra in

364 the preserved sequence (Fig. 7, 8), such that the anteriormost articulation (between the last

365 cervical and the first dorsal vertebrae) is significantly broader transversely and taller

366 dorsoventrally than the posteriormost (between dorsal vertebrae IV and V). The length of

367 each centrum varies along the column, although the fragmented, yet articulated, nature of the

368 series precludes accurate measurement in many cases (Fig. 4–8). Dorsal vertebra II appears to

369 be shorter than either dorsal vertebra I or III (Table 1), and it is presumed that, when

370 complete, the posteriormost cervical vertebra would have been anteroposteriorly longer than

371 the first dorsal vertebra.

372 The parapophysis is located on the lateral surface of the centrum, anterior to the

373 lateral pneumatic foramen, on dorsal vertebrae I and II; however, in dorsal vertebrae III–V,

374 the parapophysis is situated entirely on the neural arch (Fig. 5, 6). This pattern of

375 parapophyseal migration is consistent with the trend among transitional cervicodorsal

376 vertebrae of sauropodomorphs in general (Upchurch et al. 2004). The shape and orientation

377 of the pneumatic foramen varies with the position of the parapophysis: where the

378 parapophysis is located on the centrum, the long axis of the pneumatic foramen is inclined

379 anterodorsally–posteroventrally, whereas where the parapophysis is located entirely on the

380 neural arch, the long axis of the pneumatic foramen is aligned anteroposteriorly (Fig. 5, 6).

381 Posteriormost cervical vertebra (“specimen c1”). The posterior cotyle of this cervical

382 vertebra is preserved in articulation with the anterior condyle of dorsal vertebra I. The

383 articulation is very slightly offset (Fig. 7, 8), as if the base of the neck of Austrosaurus was

384 turned slightly to the left when the carcass was buried. The cotyle is wider transversely than it

385 is tall dorsoventrally, as is often the case in sauropod posterior cervical vertebrae (Mannion et

386 al. 2013). The preserved internal texture is camellate.

387 ===PLEASE INSERT FIGURE 5===

388 Dorsal vertebra I (“specimens c2/b1 + b4”). Longman (1933, p. 136) incorrectly

389 considered this vertebra and the preceding one to have been “from near the sacral region”,

390 based on the fact that the centra are wider transversely than they are tall dorsoventrally. He

391 also suggested that these vertebrae had been distorted dorsoventrally to such a degree that the

392 matrix infilling the neural canal had been forced into the centrum; this is incorrect. The

393 vertebra is largely undistorted—the neural canal was naturally set low in the centrum, such

394 that the dorsal margins of the condyle and cotyle were shallowly concave (and therefore

395 somewhat “heart-shaped” in anterior and posterior views) to accommodate the passage of the

396 spinal cord.

397 The anterior portion of dorsal vertebra I is preserved in articulation with the preceding

398 (cervical) vertebra, the cotyle of which obscures much of the succeeding anterior condyle.

399 The centrum is opisthocoelous, and the anterior condyle is offset from the main body of the

400 centrum by a pronounced rim around the lateral and ventral margins. This rim concomitantly

401 forms the anterior margin of the ventral fossae (see below) and the anterior border of the

402 lateral accessory foramina, close to where it presumably contacted (or even partially

403 supported) the base of the parapophysis.

404 The anterior half of the ventral surface, posterior to the condylar rim, is occupied by a

405 pair of fossae separated by a broad, anteroposteriorly-oriented ridge (Fig. 8). This ridge is

406 truncated posteriorly by the breakage of the vertebra; nevertheless, both fossae are more or

407 less completely preserved. The left fossa is deeper and larger than the right one. The

408 anteroventral margin of the vertebra (i.e. the condyle) is expressed as a shelf that forms the

409 anterior margins of the ventral fossae. Posterior to the fossae, the ventral surface of the

410 centrum is transversely convex, as the aforementioned ridge expands transversely and flattens

411 out, merging smoothly with the external surface of the posterior cotyle.

412 The majority of the external bone on the right lateral surface has been weathered away

413 (Fig. 6), with the ventrolateral margin the only portion preserving the original external

414 surface. Despite this, it is likely that only a thin veneer of superficial bone is missing given

415 that the morphology of the preserved right lateral surface approximates that of a completely

416 intact vertebral surface.

417 Two pneumatic features are present on the right lateral surface. The larger of the two

418 is the elliptical lateral pneumatic foramen (90 mm × 40 mm), which is inclined

419 anterodorsally–posteroventrally. The ventral surface of the lateral pneumatic foramen is

420 bounded by a ridge (30 mm tall dorsoventrally). The second, smaller pneumatic feature is an

421 accessory foramen (45 mm × 30 mm) that lies approximately 20 mm anterior to the lateral

422 pneumatic foramen (Fig. 6). This structure is ovate, with its long axis aligned dorsoventrally,

423 and bears surficial cortical bone around almost its entire circumference. Although it is not

424 possible to quantify its total depth, it appears to have penetrated quite deeply posteromedially

425 into the centrum.

426 The dorsal and posterior margins of the left lateral pneumatic foramen have been

427 weathered away (Fig. 5). Consequently, the preserved margins can only be measured at a

428 deeply inset level, such that the dimensions obtained (57 mm × 21 mm) understate its original

429 size. The left lateral pneumatic foramen is not entirely visible in lateral view; it is best viewed

430 in oblique posterior aspect without the posterior portion of the vertebra (“specimen b1”)

431 attached (Fig. 4 A2). This foramen is set within a fossa, and has an arc-like shape (with the

432 concave margin of the arc facing posterodorsally). The anterodorsal margin of the pneumatic

433 foramen, posteriorly adjacent to the parapophysis, penetrates most deeply. The left accessory

434 foramen (which is visible in dorsal view; Fig. 7) is situated approximately 60 mm anterior to

435 the lateral pneumatic foramen. It is ovate (45 mm dorsoventrally × 35 mm anteroposteriorly),

436 with a well-defined posterior margin (preserved with external bone) and relatively indistinct

437 anterior and dorsal margins.

438 The lateral accessory foramina preserved on both sides of this vertebra were not

439 observed on any of the subsequent dorsal vertebrae. However, the possibility that lateral

440 accessory foramina were present on the cervical vertebrae cannot be ruled out at present.

441 The bases of the parapophyses are incomplete, truncated by erosion (Fig. 5, 6). We

442 infer that each occupied a position between the two lateral foramina present on each side of

443 the centrum; it is possible that the division between the lateral pneumatic and accessory

444 foramina formed a buttress extending from the body of the centrum to the parapophysis.

445 In dorsal view, very little exterior cortical bone is visible (Fig. 7). The neural arch has

446 been almost altogether lost, fully revealing the hemispherical condyle–cotyle articulation

447 between this and the preceding cervical vertebra, as well as the margins of the neural canal of

448 this vertebra. The neural canal, represented by sedimentary matrix infill, is taller than wide.

449 Posteriorly, the preserved portion of the neural canal on “specimen c2” connects

450 (imperfectly) to the portion of the first dorsal vertebra herein labelled “specimen b4”

451 (Appendix 1), which represents the matrix between the first (“specimens c2/b1”) and second

452 dorsal vertebrae (“specimens b2/a1”).

453 In posterior view, only camellate internal bone is visible on “specimen c2” (Figure 4

454 A2). The neural canal and the internal penetrations of the pneumatic foramina are the most

455 prominent features. The posterior truncation of this portion of the vertebra highlights the

456 asymmetry of the ventral fossae (since the ventral ridge is visible in cross-section) and of the

457 placement of the neural canal opening.

458 Dorsal vertebra II (“specimen b2–4/a1”). The anterior portion of the second dorsal

459 vertebra forms part of Longman’s (1933) “specimen B”, whereas the posterior portion forms

460 part of his “specimen A”. Longman did not recognise the connection between these

461 specimens, and in fact stated that they were not consecutive (Longman 1933, p. 136). The

462 presence of a relatively ventrally positioned parapophysis on the centrum immediately

463 anterior to the pneumatic foramen—but not at the anteroventral margin—is indicative of

464 dorsal migration of the parapophysis relative to the preceding vertebra (Fig. 5, 6).

465 The anterior condyle is mostly obscured since it is articulated with the cotyle of the

466 preceding vertebra. However, CT scans of this specimen demonstrate unequivocally that it is

467 strongly opisthocoelous. The ventral surface of the centrum is transversely convex and

468 anteroposteriorly concave, with a sagittal ridge (expressed more strongly anteriorly than

469 posteriorly) as per the condition of the first dorsal vertebra (Fig. 8). The ventral fossae are

470 extremely shallow, with each situated posterior to the annulus of the anterior condyle and

471 lateral to the sagittal ridge. The right fossa is marginally deeper than the left, and is more

472 clearly demarcated, as in dorsal vertebra I. However, the lateral margins of these fossae do

473 not form strong ventrolateral ridges, distinguishing them from those present on the first dorsal

474 vertebra.

475 In lateral view, the pneumatic foramen is situated at approximately the mid-length of

476 the centrum (Fig. 5, 6). The surface of the centrum ventral to the pneumatic foramen is

477 effectively flat and faces ventrolaterally. The lateral surface of the centrum is concave

478 anteroposteriorly, as a consequence of the terminal flaring of the condyle and cotyle. The

479 parapophysis is situated immediately posterior to the condylar annulus, and anterior to the

480 pneumatic foramen. It is also broken at its base, revealing a camellate internal structure. At its

481 truncated base, the parapophysis is relatively more elongated dorsoventrally than

482 anteroposteriorly.

483 The lateral pneumatic foramen is semicircular in outline, with a slightly convex

484 posterodorsal margin. Its long axis is oriented anterodorsally–posteroventrally (length: 65

485 mm), whereas its maximum dorsoventral height is 45 mm. The foramen is dorsoventrally

486 deepest anteriorly and is posteriorly acuminate. Unlike the first dorsal, the posterior margin

487 of the pneumatic foramen does not dissipate gradually; instead, it clearly terminates 80 mm

488 anterior to the margin of the posterior cotyle.

489 “Specimen b3”, a fragment of matrix including a partial rib portion, keys into the left

490 lateral pneumatic foramen of this vertebral centrum. The rib fragment houses a large (65 mm

491 long) coel, infilled with calcite, which runs parallel to its long axis. A second smaller coel

492 occurs slightly more distally. Another extraneous fragment occurs at the left dorsolateral

493 margin of the cotyle. This fragment, which is approximately 95 mm wide, has a very compact

494 and dense internal texture and might represent another rib portion.

495 The exposed neural arch, in anterior view (Fig. 4), presents the circumference of the

496 neural canal, flanked by the pedicels. These presumably extended dorsally to form

497 centroprezygapophyseal laminae (CPRLs); indeed, close to the medial edge of each pedicel, a

498 slightly convex ridge is present which probably represents the CPRL base. The anterior

499 neural canal opening is set within a broad, shallow fossa.

500 ===PLEASE INSERT FIGURE 6===

501 Dorsal vertebra III (“specimens a2/h/e1”). The anterior portion of this vertebra, as well as

502 the posterior portion of that preceding it, was described in detail by Longman (1933) as

503 “specimen A”. Longman correctly recognised a portion of a rib preserved on the left side of

504 the specimen, situated between the two vertebrae (Fig. 5).

505 Ventrally, the bony exterior of the condylar region has been damaged, revealing the

506 pattern of the internal pneumatic coels (Fig. 8). The remaining preserved ventral surface is

507 shallowly concave transversely and anteroposteriorly. Anteriorly, the internal coels are

508 mediolaterally narrow and anteroposteriorly elongate, with the majority being approximately

509 4–10 mm wide and 15–25 mm long. Posteriorly, away from the anteriormost margin of the

510 condyle, the coels seem to become anteroposteriorly shorter and more rounded in ventral

511 profile. Further posteriorly still, the coels are not visible due to the presence of external bone.

512 Ventrolateral ridges are only very weakly defined.

513 The left lateral surface (Fig. 5) is more intact than the right (Fig. 6), although it has

514 been painted with consolidant (shellac), which has obscured some finer scale features. More

515 of the condyle is exposed on the left side, including the full extent of the condylar rim,

516 presumably due to the loss of part of the posterior cotyle of the preceding vertebra (Fig. 5).

517 On both sides of the centrum, the lateral surface ventral to the pneumatic foramen (which

518 extends for 80 mm dorsoventrally on the right side and 75 mm on the left side) is concave

519 anteroposteriorly and very slightly convex dorsoventrally; on the left side, this convexity is

520 asymmetrical, with its apex closer to the pneumatic foramen than the ventral margin.

521 The right lateral pneumatic foramen is anteroposteriorly elongate and posteriorly

522 acuminate, with rounded anterior and dorsal margins and a straighter ventral margin (Fig. 6).

523 The preserved internal dimensions are 55 mm × 37 mm, whereas the maximum external

524 dimensions are estimated to be 80 mm × 80 mm. The anterior external margin of the right

525 pneumatic foramen is situated 40 mm from the posterior margin of the cotyle of the

526 preceding vertebra.

527 The left pneumatic foramen is both internally and externally defined, although the

528 external definition is incomplete posteriorly, and the anterior margin merges imperceptibly

529 with the annulus of the anterior condyle (Fig. 5). The maximum height of the left lateral fossa

530 is 90 mm (measured externally, anterior to the mid-length), whereas the anteroposterior

531 length can only be estimated at 110 mm. The left internal pneumatic foramen has an ovate

532 outline (72 mm anteroposteriorly × 46 mm dorsoventrally), somewhat rounded anterodorsally

533 and gently convex ventrally, and seems to have an acute—albeit not acuminate—posterior

534 terminus. The total depth of the pneumatic foramen cannot be determined due to the presence

535 of matrix; nevertheless a vertical partition can be observed within the foramen (at the mid-

536 length), which has been slightly over-prepared and consequently damaged.

537 On the left lateral side, dorsal to the pneumatic foramen, the preserved neural arch has

538 been painted over (Fig. 5). The left CPRL appears to become narrower towards the

539 prezygapophysis; however, it is broken on its lateral surface, meaning that its true

540 anteroposterior thickness cannot be ascertained. Posterior to the CPRL, the lateral surface of

541 the neural arch is deeply embayed. This embayment is bounded posteriorly by a low ACDL.

542 A third posteroventrally–anterodorsally oriented lamina, presumably the PCDL, arises

543 approximately 120 mm posterior to the base of the CPRL. Although broken, the PCDL

544 appears to intersect the ACDL, approximately 60 mm dorsal to the ventral margin of the

545 embayment [positionally equivalent to a centrodiapophyseal fossa (CDF; Wilson et al.

546 2011b)], which is also the level from which both laminae originate. The diapophysis was

547 presumably located dorsal to this intersection.

548 The external bone on the right half of the neural arch has been mostly lost (Fig. 6),

549 with the exception of the anterior surface of the CPRL, which is smoothly mediolaterally

550 convex lateral to the neural canal. The shallow CDF described for the left side (see above)

551 appears to be replicated on the right side, although this can only be inferred on the basis of

552 the morphology of the preserved sub-surficial bone.

553 Anteriorly, the bases of the CPRLs extend near-vertically at about 80° to the

554 horizontal. The surface of the left CPRL base is flat, whereas the right is gently convex. The

555 minimum mediolateral widths of the CPRLs are 70 mm (left) and 60 mm (right); these

556 measurements were taken at a level that more or less coincides with the ventral surface of the

557 neural canal (i.e. the likely position of the neurocentral juncture). Dorsal to this, the CPRLs

558 expand mediolaterally; however, above the neural canal they are too incomplete to allow

559 further observation. The outline of the anterior opening of the neural canal is circular (52 mm

560 dorsoventrally × 50 mm transversely). A broad sulcus occupies the space on the neural arch

561 dorsal to the neural canal and between the two CPRLs. The posterior neural canal opening is

562 ovate (45 mm transversely × 33 mm dorsoventrally).

563 ===PLEASE INSERT FIGURE 7===

564 Dorsal vertebra IV (“specimens e2/d/f1”). Dorsal vertebra IV is preserved in three pieces,

565 although the majority of the specimen is confined to the anterior two portions. The condyle of

566 this vertebra is mostly concealed by the cotyle of the preceding vertebra, although some of

567 the condylar rim is visible, especially in ventral (Fig. 8) and right lateral views (Fig. 6).

568 The ventral surface is both transversely and anteroposteriorly shallowly concave

569 between the condyle and cotyle (Fig. 8). The inflection point of the concavity is situated

570 posterior to the mid-length. The concavity is approximately 150 mm long anteroposteriorly,

571 whereas its minimum transverse width is 100 mm, measured at mid-centrum.

572 The left lateral surface of the centrum, ventral to the pneumatic foramen, is shallowly

573 concave anteroposteriorly and flat dorsoventrally (Fig. 5). The minimum distance between

574 the ventral margin of the centrum and the ventral margin of the pneumatic foramen is 60 mm

575 internally and 65 mm externally. In left lateral view, much surficial bone remains intact,

576 although matrix obscures the left lateral pneumatic foramen. The outermost surface of this

577 fossa infill appears to preserve a sliver of surficial bone, presumably a shard of a rib shaft

578 (distal to the rib head), which might have articulated in vivo with the succeeding vertebra

579 (Fig. 5). Despite the persistence of matrix within the left pneumatic foramen, its form and

580 depth can be gauged at the anterior break between “specimen e” and “specimen d” (i.e., more

581 or less halfway through the vertebra; Fig. 4). The bilateral foramina almost meet internally

582 and are separated only by a thin septum (20 mm wide). The left lateral pneumatic foramen

583 projects 80 mm internally. In anterior cross-sectional view (Fig. 4), the foramen at depth has

584 a horizontal ventral surface and a dorsally curved upper interior surface (similar in profile to

585 that of the preceding dorsal—see Longman 1933, fig. 3). The external anteroposterior length

586 of the left lateral pneumatic foramen was no more than 120 mm.

587 In right lateral view, the lateral surface ventral to the pneumatic foramen is

588 dorsoventrally convex (~60 mm tall; Fig. 6). The right lateral pneumatic foramen is partially

589 filled with matrix, mostly in its posterior half (the anterior half being shallowly exposed),

590 although it was clearly elliptical with its long axis horizontal. The external margins of the

591 pneumatic fossa (115 mm long anteroposteriorly) blend imperceptibly with the outer centrum

592 wall and are poorly preserved anteriorly.

593 The neural arch preserves little surficial bone on all surfaces except for the ventral

594 portion of the left lateral side (Fig. 5). Here, three buttress-like laminae originate

595 approximately 150 mm dorsal to the ventral margin of the centrum. The anteriormost of these

596 laminae, which is interpreted as a shared ACPL-PCPL base, is incompletely and poorly

597 preserved but appears to be directed anterodorsally. The posterior margin of this shared

598 ACPL-PCPL base is confluent with the base of the ACDL (which is directed

599 posterodorsally), which in turn merges dorsally with the anterodorsally directed PCDL. The

600 PCDL and ACDL define a shallow, dorsally tapering, triangular centrodiapophyseal fossa

601 (CDF). Between the ACPL and ACDL, and anterior to the junction between the ACDL and

602 PCDL, another anterodorsally–posteroventrally inclined lamina is present. This is interpreted

603 as a second (stranded) PCPL, ventrally truncated by the ACDL. Between this PCPL and the

604 conjoined ACPL-PCPL mentioned above, an anterodorsally–posteroventrally inclined,

605 elongate posterior centroparapophyseal lamina fossa (PCPL-F) is present.

606 The posterior neural canal opening is sub-circular (47 mm dorsoventrally × 42 mm

607 transversely). The surficial bone enveloping the bases of the centropostzygapophyseal

608 laminae (CPOLs) is not preserved, although there are indications that the bases were

609 approximately 55–60 mm wide mediolaterally. The medial margins of the CPOLs merge

610 smoothly with the ventral margin of the posterior neural canal opening.

611 A large fragment of dorsal rib is preserved at the interface between dorsal vertebrae

612 III and IV, with its long axis aligned dorsoventrally (Fig. 5). Despite its preserved position,

613 the rib most probably relates to dorsal vertebra III. The preserved proximodistal length of the

614 fragment is 175 mm and its internal structure is discernible, belying the presence of two large

615 (30–40 mm long) coels, in addition to five smaller (<15 mm long) coels. Disregarding these

616 coels, the internal structure of the rib is spongy towards the proximal end and much denser

617 distally, such that it is almost solid.

618 ===PLEASE INSERT FIGURE 8===

619 Dorsal vertebra V (“specimens f2 + k”). This vertebra comprises the anterior half to two-

620 thirds of the centrum, in addition to an anteroposteriorly short section of the neural canal,

621 preserved as a natural internal cast. The maximum preserved anteroposterior length of the

622 specimen is 210 mm, whereas the maximum preserved transverse width of the anterior

623 condyle is 240 mm. Surficial bone is only preserved on the condyle; the only other region of

624 this vertebra preserving near-surficial bone is the left lateral surface, ventral to the left

625 pneumatic foramen. The condyle is separated on the right side from the cotyle of dorsal

626 vertebra IV by as little as 3 mm of matrix; ventrally, the intervening matrix is 8 mm thick.

627 The ventral surface of this vertebra is too poorly preserved to allow precise

628 determination of its external morphology (Fig. 8). The exposed internal coels appear to be

629 anteroposteriorly elongate, although their lengths cannot be determined accurately because of

630 their poor marginal preservation and anastomosing nature. Generally, the mediolateral width

631 of these coels is 10 mm.

632 The left lateral surface of the centrum is gently convex dorsoventrally, and is gently

633 concave anteroposteriorly (Fig. 5). Matrix fills in and defines the shape of the left lateral

634 pneumatic foramen, which presents a lenticular outline, oriented anterodorsally–

635 posteroventrally (87 mm × 50 mm). On the right surface, no surficial bone is preserved (Fig.

636 6). However, the anteroventral half of the border of the external pneumatic foramen appears

637 to be delineated, whereas the posterodorsal half is not.

638 The neural canal is represented by a short section of natural cast which is circular

639 anteriorly (52 mm transversely × 50 mm dorsoventrally) and ovate posteriorly (46 mm

640 transversely × ~40 mm dorsoventrally). The maximum length of the preserved natural neural

641 canal cast is 75 mm.

642 Additional, positionally indeterminate vertebral specimens. A number of additional

643 fragments smaller than the blocks comprising the main sequence of articulated vertebrae are

644 present. Some of these might derive from dorsal vertebrae more posterior than dorsal vertebra

645 V. Although only a few of these fragments are described below (others are too poorly

646 preserved to be informative), a full listing is provided in the Appendix.

647 A fragment of dorsal vertebra (“specimen f”) preserves the right posteroventral

648 portion of a centrum, including the posteroventral margin of the pneumatic foramen, the base

649 of the posterior cotyle (which has mostly been lost to erosion), and matrix representing the

650 position of the intervertebral disc. The internal camellate texture is clearly visible. The

651 ventral surface was very shallowly concave anteroposteriorly and transversely, whereas the

652 ventral portion of the lateral surface was dorsoventrally convex ventral to the pneumatic

653 foramen; anteroposteriorly it was presumably flat. The long axes of the internal pneumatic

654 coels run parallel to the cotylar margin, arranged in two concentric rings (ventrally) and four

655 concentric rings (dorsally). Anterior to the cotylar margin, the coels become more elongate

656 anteroposteriorly, more irregularly distributed, and less palisade-like in arrangement.

657 One narrow fragment comprises two thin slivers of possible vertebral centra with

658 intervening sediment (“specimen q”). These vertebral portions could conceivably represent

659 an articulation between two dorsal centra, implying that more than six vertebrae of

660 Austrosaurus were originally preserved; this, however, is speculative. The concave element is

661 extremely incompletely preserved, represented mostly by dense surface bone which forms a

662 thin veneer; the convex element is more substantially represented. In cross-section, the

663 surface is formed by a 3–5 mm thick section of dense bone, which is followed posteriorly by

664 a palisade of internal coels. These coels are anteroposteriorly elongate (up to 37 mm

665 anteroposteriorly × 16 mm mediolaterally; most are smaller), and are separated from their

666 neighbours by 1–2 mm thin septa. The intervertebral matrix is between 10–20 mm thick and

667 varies from one side of the specimen to the other.

668 A 120 mm deep fragment comprising the interface between two articular units and the

669 intervening sediment (“specimen p”) seems to represent an articulated set of zygapophyses.

670 The majority of the external surface has been lost; consequently, our interpretation is

671 tentatively based on the morphology of the subsurface, which is presumed to replicate the

672 original external morphology. The larger of the two preserved zygapophyses appears to

673 represent a right postzygapophysis, embayed medially by an arc representing the right half of

674 the spinopostzygapophyseal fossa (SPOF). This latter, medially facing surface retains a small

675 section of external bone. Overall, the posterior/posterolateral surface of the postzygapophysis

676 is convex, with the dorsal-most preserved portion being slightly more ridge-like. This

677 morphology is consistent in general with sauropod postzygapophyses extending dorsally as

678 narrowed spinopostzygapophyseal laminae. Thus, the more convex dorsal part of “specimen

679 p” likely represents the base of that lamina. The counterpart right prezygapophysis is

680 comparatively incomplete, but is separated from the postzygapophysis by a plane of sediment

681 that is 10–20 mm thick, which likely conforms to an in vivo inter-articular gap. It appears that

682 the internal coels of the postzygapophysis formed a palisade with their long axes

683 perpendicular to the inter-zygapophyseal gap (the coels are about 20 mm dorsoventrally × 9

684 mm mediolaterally). Dorsal to the gap, the coels are larger in size (35 mm × 15 mm) and bear

685 anteroposteriorly longer axes. The internal coel patterning is less well-defined for the

686 prezygapophysis; there is a dense 5–10 mm thick layer of tissue at the prezygapophyseal

687 articular facet, which thins medially.

688 ===PLEASE INSERT FIGURE 9===

689 Dorsal ribs. A total of six dorsal ribs are preserved in the Austrosaurus hypodigm (Fig. 9).

690 Five of these are from the left side, representing dorsal ribs I–V, whereas the other is

691 interpreted to be right dorsal rib IV, based on the congruence between its morphology and

692 that of left dorsal rib IV. The anterior four ribs are fairly consistent in their morphology. The

693 proximal ends are thickened, subtriangular in cross-section, and each has a somewhat

694 concave anteromedial margin near the point at which the capitulum and tuberculum would

695 have met. In contrast, the distal ends (of dorsal ribs I–III, at least) are plank-like, being much

696 longer anteroposteriorly than wide mediolaterally (anteroposterior:mediolateral ratio in dorsal

697 rib I = 6.13; dorsal rib II = 4.87; and dorsal rib III = 3.14). The lateral surfaces of the ribs are

698 flat, whereas the medial surfaces are convex. Distal to the rib head, each of the first four ribs

699 bears a longitudinal groove on the posterior surface, which extends halfway down the shaft.

700 In cross-section, each of the anterior four ribs is mediolaterally broadest in its posterior third,

701 and the anterior margin of each is tapered. The portion of each rib bearing a posterior groove

702 is arrowhead-shaped in cross-section, whereas distal to the groove the cross-section is “D”-

703 shaped, with the lateral margin being straight. Dorsal rib V does not have a plank-like distal

704 end, and also lacks the well-defined posterior groove present in dorsal ribs I–IV. At mid-

705 length, the cross-section of this rib is almost circular. Distally, dorsal rib V is mediolaterally

706 compressed but not particularly elongate anteroposteriorly.

707 The shafts of all six ribs are well preserved, whereas the proximal ends are not. The

708 structure of the internal bone of several of the anterior dorsal ribs demonstrates that the

709 proximal ends were pneumatised, a feature which can also be seen in some of the fragments

710 collected in the 1930s. This pneumatisation presumably contributed to the poor preservation

711 of the proximal ends.

712 Discussion

713 The taphonomy of the Austrosaurus mckillopi type series

714 ===PLEASE INSERT FIGURE 10===

715 Generally, it is quite unusual to find sauropod skeletal remains in marine settings

716 (Mannion & Upchurch 2010) because they were fully terrestrial animals. It is tempting to

717 correlate the frequency of preservation of terrestrial vertebrate fossils in a marine setting with

718 land area proximity. However, Buffetaut (1994) warned against drawing this conclusion too

719 readily, noting that some workers had suggested that dinosaur carcasses could have been

720 carried for hundreds or thousands of kilometres (Martill 1988).

721 Given that the type specimen of the terrestrial sauropod Austrosaurus mckillopi was

722 found in the marine Allaru Mudstone, it is clearly allochthonous. As outlined above, QM

723 F2316 constitutes an articulated sequence of vertebrae with ribs (KK F1020; Fig. 10);

724 consequently, we were able to utilise the skeletal articulation and completeness metrics

725 proposed by Beardmore et al. (2012) to elucidate its taphonomic history. Beardmore and

726 colleagues’ system separates the skeleton into nine skeletal units (head, neck, trunk, tail, ribs,

727 left and right forelimbs + pectoral girdles, left and right hind limbs + pelvic girdles) and ranks

728 articulation and completeness on a scale of 0–4, where 0 indicates low completeness and

729 disarticulation, and 4 indicates high completeness and full articulation. We also adopted the

730 classification system proposed by Syme & Salisbury (2014) to quantify the degree of

731 articulation between skeletal units, where F denotes full articulation between adjacent units, P

732 denotes partial articulation, and D denotes disarticulation.

733 The cervical and dorsal vertebrae have an articulation score of 4—fully articulated

734 with no gaps or breaks—and the position of the ribs in the field suggests a similarly high

735 articulation score. The neck and trunk skeletal units are fully articulated with one another

736 (inter-unit articulation category F), and the rib and trunk skeletal units were either semi-

737 articulated (inter-unit articulation category P) or fully articulated. Given that non-

738 titanosaurian titanosauriforms had between 13 [Giraffatitan (Janensch 1950)] and 17

739 [Euhelopus (Wilson & Upchurch 2009)] cervical vertebrae, and that titanosaurs had between

740 14 [Futalognkosaurus (Calvo et al. 2007)] and 17 [Rapetosaurus (Curry Rogers 2009)]

741 cervical vertebrae, we calculate that the completeness of the Austrosaurus neck is

742 approximately 6–7%—a completeness score of 1. If the total number of dorsal vertebrae in

743 Austrosaurus fell between the extremes known among titanosauriforms [10 dorsals in

744 Futalognkosaurus (Calvo et al. 2007), Overosaurus (Coria et al. 2013) and Trigonosaurus

745 (Campos et al. 2005), and 13 in Euhelopus (Wilson & Upchurch 2009)], the completeness

746 score would be between 38–50%—a completeness score of 2. Assuming that the number of

747 dorsal ribs was precisely double the number of dorsal vertebrae, and given that six ribs were

748 recovered, we calculate a percentage completeness ranging between 23–30% for the ribs—a

749 completeness score of 1 or 2. Thus, QM F2316 displays relatively low completeness but

750 relatively high articulation; this suggests that the carcass underwent a period of decay prior to

751 burial in a low energy environment (Beardmore et al. 2012).

752 The body of Austrosaurus most likely drifted out to sea as a consequence of a process

753 termed ‘bloat and float’ (Allison & Briggs 1991), wherein gas produced during endogenous

754 decay built up in the body tissues and the digestive tract, which caused the body to swell and

755 become positively buoyant in water (Schäfer 1972, Davis & Briggs 1998, Rogers & Kidwell

756 2007). The system of pneumatic diverticula present in sauropods (Wedel 2009), and the high

757 level of postcranial pneumatisation developed in titanosauriforms (Wilson & Sereno 1998)

758 and exemplified by titanosaurs (Cerda et al. 2012) would have affected the buoyancy of these

759 dinosaurs (Henderson 2004a) and might have prolonged the duration of the flotation stage;

760 however, the effect of this has not yet been quantified. As decay progressed and the integrity

761 of the body tissues was compromised, gases would have escaped the carcass. This, along with

762 continual saturation of body tissues, would have resulted in the carcass becoming negatively

763 buoyant and caused it to sink (Fig. 11). For the last cervical vertebra, the first five dorsal

764 vertebrae, and several ribs to have remained in articulation, some connective soft tissues must

765 have been intact prior to the carcass settling on the sea floor (as suggested by Molnar 2010),

766 left side down.

767 ===PLEASE INSERT FIGURE 11===

768 It is possible that cranial and postcranial elements missing from the Austrosaurus site

769 were detached by scavengers during the ‘bloat and float’ phase. In this scenario, the

770 disruption of the integrity of the soft tissue by scavengers might have facilitated the

771 separation and sinking of a portion of soft tissue containing presacral vertebrae and ribs; the

772 remainder of the carcass would have been consumed and/or scattered. Another possibility is

773 that the carcass progressed to the ‘advanced decay stage’, unaffected by scavenging,

774 whereupon soft tissue decayed to such a degree that the carcass broke into discrete segments;

775 the posteriormost ribs appear to have been ‘stacked’, so sufficient soft tissue must have

776 decayed prior to burial to allow these elements to partially disarticulate. Given that the

777 Austrosaurus site has been exhausted, and if we assume that the entire skeleton was buried

778 and fossilised in life position laying on its left side, the missing elements must have eroded

779 during post-diagenetic subaerial weathering. Although this might be true for elements

780 positioned stratigraphically higher than the remainder of the recovered skeleton (such as the

781 right ribs), for the rest of the skeleton we find this to be the least likely scenario because of

782 our failure to discover additional skeletal material either vertically beneath, or laterally

783 throughout, the soil profile at the site. It is important to note that the presence of a single right

784 dorsal rib lying near to its in vivo position does not necessarily negate this interpretation, but

785 rather suggests only enough soft tissue decay occurred to allow this single rib to disarticulate

786 prior to burial. The lack of skeletal element transport after disarticulation also aligns with the

787 interpretation of deposition below wave-base in a calm water environment.

788 ===PLEASE INSERT FIGURE 12===

789 Given the presence of marine ‘reptile-fall’ type communities during the Mesozoic

790 (Kaim et al. 2008, Wilson et al. 2011a, Danise & Higgs 2015), it is conceivable that any

791 sauropod remains lying at the sediment-water interface would have been colonised by benthic

792 communities, forming a ‘saurian deadfall’ [sensu Hogler (1994) and Reisdorf & Wuttke

793 (2012)]. Colonisation might have taken place within weeks or months, resulting in bone

794 surface modifications and bone erosion from chemical dissolution (Trueman & Martill 2002,

795 Adams 2009, Anderson & Bell 2014). Although there is no evidence of encrusting biota or

796 eroded bone on the Austrosaurus specimen, remains of the molluscs Inoceramus sp. and

797 Beudanticeras sp. were recovered from the site and are preserved within matrix adhering to

798 the vertebrae (Fig. 12). It is possible that the bacteriophage filter-feeding Inoceramus sp. [a

799 mode of life proposed by Henderson (2004b)] were attracted to bacteria feeding on the

800 decaying Austrosaurus remains. The apparent lack of other benthic scavengers, including

801 crustaceans and teleost fish similar to those recovered from the underlying Toolebuc

802 Formation (Wilson et al. 2011a, Smith & Holland 2016), could be a result of either true

803 absence or preservation bias. There is also evidence of benthic feeding elasmosaurids and

804 protostegid turtles occupying the Eromanga Sea at this time, with fossilised mollusc-rich gut

805 contents and coprolites from both middle and upper Albian units including the Allaru

806 Mudstone (McHenry et al. 2005, Kear 2006). If plesiosaurs [e.g. the pliosaur Kronosaurus

807 (Longman 1924, 1930, White 1935, Romer & Lewis 1959, McHenry 2009, Holland 2015)]

808 scavenged carcasses when they were available, and if the remains of Austrosaurus had a

809 relatively long residence time at the sediment-water interface, they would most likely have

810 been disarticulated and possibly destroyed before burial and fossilisation could occur. That

811 the remains were preserved relatively intact suggests that this isolated portion of the

812 Austrosaurus carcass was buried after only a short period of decay at the sediment–water

813 interface during the ‘mobile scavenger stage’ or the beginning of the ‘enrichment

814 opportunistic stage’ (sensu Smith & Baco 2003, Kaim et al. 2008), before an abundant

815 saurian deadfall community could form.

816

817 The phylogenetic position of Austrosaurus mckillopi

818 Previous opinions on the phylogenetic position of Austrosaurus mckillopi. When

819 Austrosaurus was first described during the 1930s, sauropod inter-relationships in general

820 were poorly understood. Longman (1933) considered Austrosaurus to be more specialised

821 than Rhoetosaurus brownei (Longman 1926) on the basis of the more complex internal

822 vertebral structure of the former, and specifically stated that the internal vertebral structure of

823 Diplodocus carnegii (Hatcher 1901) was quite similar to that of Austrosaurus. Longman

824 (1933) also suggested that Austrosaurus bore no close relationship to the Argentinean

825 sauropods Titanosaurus (now Neuquensaurus), Laplatasaurus or Antarctosaurus (Huene

826 1929), and suggested that it was not a member of Titanosauridae. Instead, Longman (1933)

827 tentatively suggested that Austrosaurus was an advanced member of the Cetiosauridae.

828 Coombs & Molnar (1981) were undecided on the phylogenetic placement of

829 Austrosaurus, suggesting that classification as ‘Sauropoda incertae sedis’ was advisable if not

830 satisfactory. In a non-technical summary of Austrosaurus, Thulborn (1987, pp. 44–46)

831 provided a frank assessment of the then state-of-play of sauropod phylogenetics. His

832 observations of the type specimen allowed him to state that Austrosaurus was not allied with

833 diplodocids, camarasaurs or brachiosaurs, and that the classification of Austrosaurus as a

834 cetiosaur was essentially meaningless, since at the time Cetiosauridae was “…really a waste-

835 basket category for any generalised or primitive-looking sauropods that can’t easily be

836 classified elsewhere” (Thulborn 1987, p. 44).

837 Molnar (2001b, p. 141) suggested that Austrosaurus was a titanosaur, based on

838 characters outlined by Salgado et al. (1997). Molnar (2001b, pp. 141, 143) regarded at least

839 some of the preserved vertebrae as posterior dorsal vertebrae, and noted the presence of the

840 following characters: centroparapophyseal laminae present; pneumatic foramina eye-shaped

841 and deep; centra opisthocoelous; and “spongy” internal texture. Given that we now know that

842 most of the preserved vertebrae of Austrosaurus are actually anterior dorsal vertebrae, the

843 significance of many of these characters is greatly diminished.

844 The only phylogenetic analysis in which Austrosaurus has been included is that of

845 Upchurch et al. (2004). On the basis of this analysis, Austrosaurus was recovered as a non-

846 lithostrotian titanosaur (Upchurch et al. 2004, p. 310), an interpretation followed by Barrett &

847 Upchurch (2005, p. 153). However, as has been mentioned elsewhere (Mannion et al. 2013,

848 p. 154, Poropat et al. 2015a, pp. 92–93), the scores for this taxon relied heavily upon material

849 referred to Austrosaurus sp. by Coombs & Molnar (1981), rather than the holotype. All

850 specimens referred to Austrosaurus sp. by Coombs & Molnar (1981) were removed from

851 Austrosaurus by Molnar (2001b); one of these (QM F7292) now constitutes the holotype of

852 Wintonotitan wattsi (Hocknull et al. 2009, Poropat et al. 2015a).

853 Molnar & Salisbury (2005) discussed the phylogenetic placement of Austrosaurus in

854 their revision of Australian Cretaceous sauropods. They noted the presence of three vertebral

855 laminae on Longman’s (1933) Specimen A (here, dorsal vertebra III): the anterior

856 centroparapophyseal lamina (ACPL), anterior centrodiapophyseal lamina (ACDL), and

857 posterior centrodiapophyseal lamina (PCDL). More importantly, however, Molnar &

858 Salisbury (2005) also incorporated some of the material not described by Longman (1933) in

859 their assessment of Austrosaurus. Of especial note was their consideration of what has now

860 been recognised as the middle section of dorsal vertebra IV (“specimen d”; see Appendix). In

861 the text, Molnar & Salisbury (2005, pp. 456–457) state that “…it is unclear which end of the

862 specimen is anterior”, although they appear to contradict this statement by stating that “the

863 right side is seen in Figure 20.1”. However, if the laminae labels on their figure are taken at

864 face value, then Molnar & Salisbury (2005) depicted the specimen in question in what can

865 only be interpreted as left lateral view. On the basis of the identification of five laminae

866 (ACDL, PCDL, prezygodiapophyseal lamina [PRDL], spinodiapophyseal lamina [SPDL],

867 and postzygodiapophyseal lamina [PODL]), they suggested that Austrosaurus was a

868 titanosaur; however, they did not specify which, if any, of these laminae lent support to this

869 claim. Based on our observations, the majority of these laminae were either misidentified

870 (PRDL, SPDL, PODL) and cannot be observed, or were incorrectly positioned (PCDL) by

871 Molnar & Salisbury (2005).

872 Hocknull et al. (2009, p. 40) considered Austrosaurus as a nomen dubium because

873 they regarded the holotype specimen as inadequate for diagnostic purposes; they also

874 suggested that a neotype should be designated. In contrast, Agnolin et al. (2010, p. 262)

875 regarded Austrosaurus as a valid taxon, assigning it to Titanosauriformes. Molnar (2011a), in

876 another discussion of the holotype specimen of Austrosaurus, noted a personal

877 communication from Zhao Xijin who “suggested that some of the vertebrae might be

878 posterior cervicals” (Molnar 2011a, p. 322), a proposal with which we agree. Molnar stated

879 that none of the preserved pieces of dorsal rib were pneumatised; this led him to suggest that

880 Austrosaurus was a non-titanosauriform sauropod, since Wilson & Sereno (1998) and Wilson

881 (2002) had earlier identified ribs with proximal pneumatic chambers as being synapomorphic

882 of this clade. We disagree with the removal of Austrosaurus from Titanosauriformes on this

883 basis: one specimen catalogued as part of the Austrosaurus type series appears to be close to

884 the proximal end of a dorsal rib, and it is pneumatised, as are all of the rib proximal ends

885 recovered in 2014–2015. One non-pneumatised rib portion is embedded in matrix in

886 association with the articulated dorsal vertebrae of the type series; however, this rib segment

887 represents a region somewhat distal to the proximal end, and would not have been expected

888 to be pneumatised.

889 Mannion & Calvo (2011) tentatively agreed with the designation of Austrosaurus as a

890 nomen dubium by Hocknull et al. (2009), but also assigned it with reservations to

891 Titanosauria. In contrast, Mannion et al. (2013, p. 154) unequivocally regarded Austrosaurus

892 as a nomen dubium, pending restudy, and considered it to be classifiable only as far as

893 Titanosauriformes. Poropat et al. (2015a) also regarded Austrosaurus was a nomen dubium,

894 pending reappraisal.

895

896 Phylogenetic distribution of key anatomical features. Some of the features observed in this

897 study have potential significance for the phylogenetic position of Austrosaurus mckillopi.

898 These include the ventral ridges on the anterior dorsal vertebrae, the internal pneumatic

899 features of the vertebrae, and the morphology of the dorsal ribs.

900 (1) Ventral ridges on anterior dorsal vertebrae. The ventral surfaces of dorsal vertebrae I

901 and II in Austrosaurus each bear prominent midline ridges, bounded on each side by a

902 shallow fossa. None of the other dorsal vertebrae preserved appear to have had ventral ridges,

903 although it is possible that ridges were present on vertebrae missing from the type series.

904 Although several titanosauriform taxa have ventral ridges in their middle–posterior

905 dorsal vertebrae (e.g. the brachiosaurids Brachiosaurus and Giraffatitan; Upchurch 1998), far

906 fewer possess these in their anterior dorsal vertebrae (Mannion et al. 2013, Poropat et al.

907 2016). In Euhelopus, a euhelopodid somphospondylan (D'Emic 2012), one of the vertebrae at

908 the cervicodorsal transition (possibly the first dorsal, but probably the last cervical) bears a

909 ventral median ridge set within a concavity (Wilson & Upchurch 2009, pp. 212, 219). This

910 contrasts with the ventral ridges on the first two dorsals of Austrosaurus, which are not set

911 within concavities.

912 Phuwiangosaurus is commonly resolved as a non-titanosaurian somphospondylan

913 (Suteethorn et al. 2010, Carballido et al. 2011b, 2012b, D'Emic 2012, 2013, Mannion et al.

914 2013, Carballido & Sander 2014, Lacovara et al. 2014, Poropat et al. 2015b, Upchurch et al.

915 2015, Gorscak & O'Connor 2016, Poropat et al. 2016) or a basal titanosaur (Upchurch et al.

916 2004, Carballido et al. 2011a, 2015) in phylogenetic analyses. The description of a ventral

917 ridge on an anterior dorsal vertebra (Martin et al. 1999, p. 47) in this taxon matches the

918 morphology of the ventral ridges of Austrosaurus. In more posterior dorsal vertebrae, the

919 ventral ridge was only faintly expressed (Martin et al. 1999, p. 49), whereas in a referred

920 specimen of Phuwiangosaurus, it was determined that ridges were present on dorsal

921 vertebrae III–VII (Suteethorn et al. 2009), demonstrating some intraspecific variation.

922 Relatively few titanosaurs are reported to have had ventral ridges on their anterior

923 dorsal vertebrae. Exceptions include Barrosasaurus (Salgado & Coria 2009), Overosaurus

924 (Coria et al. 2013) and Lirainosaurus (Díez Díaz et al. 2013), although perhaps most notable

925 is Opisthocoelicaudia in which ventral ridges set within deep fossae are present on all dorsal

926 vertebrae (Borsuk-Białynicka 1977). Among these titanosaurs, only the ventral ridges of

927 Lirainosaurus approximate those of Austrosaurus in both morphology and distribution.

928 As the above summary suggests, the presence of a ventral ridge that is not set within a

929 fossa in the anterior dorsal vertebrae of Austrosaurus mckillopi is unusual among

930 Titanosauriformes, but not autapomorphic, given the presence of similar ventral ridges in the

931 anterior dorsal vertebrae of Phuwiangosaurus and Lirainosaurus. Ventral ridges are not

932 known in the anterior dorsal vertebrae of any other Australian sauropod, although this might

933 simply be because no other Australian sauropod specimens described to date preserve dorsal

934 vertebrae I or II. Additional specimens will be necessary to determine how widespread this

935 feature was among Australian sauropods, if it was present in taxa other than Austrosaurus at

936 all.

937 ===PLEASE INSERT FIGURE 13===

938 (2) Internal texture of the vertebrae. The pneumatic nature of sauropod vertebrae, first

939 observed by Seeley (1870), was also considered by Longman (1933) in his description of

940 Austrosaurus mckillopi. However, Longman did not think that the internal cavities of the

941 vertebrae connected with the pneumatic foramina: he instead tentatively supported Owen’s

942 (1876) suggestion that the internal spaces had been filled with “chondrine”. The basis for this

943 contention was that the chemical composition of the matrix within the coels was different

944 from that surrounding the exterior of the specimen. Later, however, Longman changed his

945 mind, possibly influenced by Janensch’s (1947) work: in a 1949 newspaper article on

946 Austrosaurus, he stated, “…the body of the vertebrae is composed of a multitude of small

947 chambers, and it is also hollowed out on each side. Probably these chambers were filled with

948 air in life.” (Longman 1949, p. 2).

949 Wedel et al. (2000) determined that titanosauriforms show three types of internal

950 texture in their presacral vertebrae: semicamellate, camellate, and somphospondylous. In the

951 case of Austrosaurus, the preserved portions of the last cervical and first five dorsal vertebrae

952 are camellate. Although it is probable that the presacral vertebrae of Austrosaurus were

953 somphospondylous [following the criteria of Wedel et al. (2000) and Wedel (2003)], the

954 incomplete preservation of the neural arches and the absence of neural spines prevent this

955 from being demonstrated with certainty. Subsequently, Wedel (2003) highlighted the fact that

956 the internal texture of the vertebrae of any given sauropod will show variation within the

957 column, specifically stating that the internal texture of the posterior cervical vertebrae tended

958 to be the most complex.

959 In order to further assess the internal texture of the Austrosaurus mckillopi type series,

960 all specimens catalogued as QM F2316 were CT scanned at Greenslopes Private Hospital,

961 Brisbane on a Siemens/Somatom Definition Flash CT scanner at 100 kV and 8 mAs, and 140

962 kV and 22 mAs, with a slice increment of 0.4 mm. Although the majority of the scans were

963 affected by significant artifacts and failed to adequately resolve the internal structure of the

964 specimens, some variation in the internal texture of each vertebra can still be seen (Fig. 13).

965 The laminae within the anterior condyles project posteromedially from the extremities

966 towards the centre of the vertebra, forming an anastomosing network. The chambers between

967 these partitions are dorsoventrally taller than they are long anteroposteriorly or (especially)

968 wide mediolaterally. Particularly notable in axial section (Fig. 13A) is the fact that the

969 camellae are most densely packed, into what appear to be three concentric layers, in the

970 posterior cotyles. Similar observations to these were made by Molnar (2011a, pp. 332–333),

971 based on inspection of the broken surfaces of the Austrosaurus type series. His recognition

972 that a densely packed, concentric layer of camellae was also present in the posterior cotyle of

973 a dorsal vertebra of Saltasaurus loricatus figured by Powell (1992, fig. 16d, 2003, pl. 30d)

974 was particularly insightful. However, the CT scans have revealed that there were three

975 concentric layers of camellae in the posterior cotyles of the dorsal vertebrae of Austrosaurus,

976 whereas there is only one in those of Saltasaurus. It is probable that the reduced number of

977 concentric layers of camellae in the posterior condyles of the dorsal vertebrae of Saltasaurus

978 is a consequence of the significantly smaller size of this animal; taking this line of reasoning

979 in the other direction, it would seem safe to presume that titanosauriforms larger than

980 Austrosaurus would have had more strongly reinforced posterior cotyles, and that this

981 strengthening might have been manifested as additional concentric layers of camellae. CT

982 scans of the dorsal vertebrae of the Brazilian Cretaceous titanosaur Austroposeidon

983 magnificus show concentric lamina rings that mirror the condyle/cotyle in shape (Bandeira et

984 al. 2016), which would appear to support this hypothesis. However, Bandeira et al. (2016, p.

985 19) interpreted these laminae as “intercalated growth structures”. Studies on the internal

986 structure of the presacral vertebrae of large titanosaurs [e.g. Argentinosaurus (Bonaparte &

987 Coria 1993), Puertasaurus (Novas et al. 2005), Futalognkosaurus (Calvo et al. 2007),

988 Dreadnoughtus (Lacovara et al. 2014), Alamosaurus (Fowler & Sullivan 2011, Tykoski &

989 Fiorillo 2017)] and non-titanosaurian somphospondylans [e.g. Daxiatitan (You et al. 2008)]

990 will be needed to test this hypothesis.

991 (3) Dorsal rib morphology. The dorsal ribs of very few titanosauriforms have been

992 adequately described: normally, the presence of proximal pneumatisation, and a vague

993 allusion to plank-like or non-plank-like distal rib ends, is all that is reported. Regarding one

994 particular aspect of rib morphology in a broader taxonomic scope, few macronarian

995 sauropods have had the cross-sectional shapes of their dorsal ribs documented in any detail.

996 Exceptions include Giraffatitan brancai (Janensch 1950) and Camarasaurus sp. (Waskow &

997 Sander 2014), for which multiple rib cross-sectional outlines have been illustrated, and

998 Opisthocoelicaudia skarzynskii (Borsuk-Białynicka 1977), for which the cross-sectional

999 outline of each preserved dorsal rib was described.

1000 As noted by Waskow & Sander (2014), each rib of Camarasaurus shows significant

1001 cross-sectional shape variation along its length, and the ribs show significant variability when

1002 considered collectively. Notably, the distal ends of the dorsal ribs of Camarasaurus appear to

1003 show the plank-like morphology generally considered to be a synapomorphy of

1004 Titanosauriformes (Wilson 2002). By contrast, the cross-sections depicted by Janensch

1005 (1950) of the anterior dorsal ribs of Giraffatitan—a titanosauriform by definition—do not

1006 conform to the criteria for being considered plank-like. Opisthocoelicaudia, however,

1007 possesses the distally plank-like anterior dorsal ribs typical of titanosauriforms: Borsuk-

1008 Białynicka (1977) observed that dorsal ribs III, IV and V had flattened distal ends; that rib VI

1009 was rounded along most of its length but flattened distally; and that ribs VII to IX were

1010 rounded in cross-section. The preserved ribs of Austrosaurus show a similar pattern to those

1011 of Opisthocoelicaudia, although the transition from plank-like to rounded was evidently

1012 completed by dorsal rib V.

1013 ===PLEASE INSERT FIGURE 14===

1014 All of the dorsal ribs of the somphospondylan Astrophocaudia (D'Emic 2013), and all

1015 (except dorsal rib I) in the brachiosaurid Cedarosaurus (Tidwell et al. 1999), were described

1016 as not plank-like. Only the middle and posterior dorsal ribs of Paludititan were described as

1017 being plank-like—the anterior dorsal ribs were “…rounded oval in cross-section and not

1018 particularly flattened…” (Csiki et al. 2010, p. 304). All dorsal ribs of Malawisaurus (Gomani

1019 2005) and Rukwatitan (Gorscak et al. 2014) were described as having flattened shafts

1020 irrespective of serial position. The dorsal ribs of Epachthosaurus appear to show some

1021 similarity to those of Austrosaurus inasmuch as the anteriormost elements were described as

1022 being distally plank-like, whereas the posteriormost were cylindrical in cross-section

1023 (Martínez et al. 2004). Dorsal ribs I and II of Overosaurus (Coria et al. 2013) are not plank-

1024 like at their distal ends, and both the anterior and posterior margins of dorsal ribs II and III in

1025 Overosaurus are expanded at the proximal end.

1026 Among Australian sauropods, perfunctory comparisons can be made between those of

1027 Austrosaurus mckillopi and some of the preserved dorsal ribs of Diamantinasaurus matildae

1028 (Fig. 14; Hocknull et al. 2009, Poropat et al. 2015b). However, these are limited by the fact

1029 that the type specimen of Diamantinasaurus was not found articulated. As such, the eight

1030 Diamantinasaurus rib sections analysed herein were assigned a letter (from A–H) based on

1031 their presumed serial position (Fig. 14). Rib A appears to represent an anterior dorsal rib,

1032 possibly the anteriormost. The cross-section of this rib is crescentic—quite unlike any of

1033 those observed in Austrosaurus. It is possible that the corresponding section(s) of the

1034 anteriormost dorsal rib(s) of Austrosaurus were not preserved, hence the morphological

1035 incongruity. Rib B preserves a complete proximal end but is incomplete distally; that it

1036 appears to taper distally so rapidly might indicate that it was from the posterior half of the

1037 thorax. Rib C is morphologically congruent with dorsal rib III of Austrosaurus (Fig. 9). The

1038 ribs of Diamantinasaurus labelled D–F in Fig. 14 are presumably from the anterior half of the

1039 thorax, although their incomplete preservation makes it difficult to determine how close the

1040 portions were to the distal ends of their respective ribs. The ribs labelled G and H are

1041 interpreted to have been situated in the posterior half of the thorax based on their rounded

1042 cross-sections. Further work on the dorsal ribs of Diamantinasaurus matildae and

1043 Savannasaurus elliottorum (Poropat et al. 2016) will hopefully shed light on the variability

1044 within and between Australian Cretaceous sauropods.

1045

1046 Revised phylogenetic placement of Austrosaurus mckillopi. On the basis of the preserved

1047 remains, it can be demonstrated that Austrosaurus mckillopi is a titanosauriform sauropod.

1048 The pneumatisation of the proximal ends of the dorsal ribs (Wilson & Sereno 1998), along

1049 with the plank-like morphology of the distal ends of the anterior dorsal ribs (Wilson 2002),

1050 suggest titanosauriform affinities for Austrosaurus. The cervical and anterior dorsal vertebrae

1051 of Galveosaurus herreroi, a taxon resolved as the sister taxon to Titanosauriformes by

1052 Mannion et al. (2013), show camellate internal texture (Barco et al. 2006, Barco Rodriguez

1053 2009); however, the dorsal ribs of Galveosaurus are not pneumatised. Therefore, the

1054 combination of features presented by Austrosaurus (i.e., presacral vertebrae with camellate

1055 internal texture + dorsal ribs with pneumatised proximal ends) allows it to be placed within

1056 Titanosauriformes.

1057 Despite the augmentation of the Austrosaurus mckillopi type specimen, the remains of

1058 this sauropod are still frustratingly incomplete. Very few phylogenetic characters can be

1059 scored, and even fewer can be scored without reversion to estimation or approximation.

1060 Consequently, it is difficult to determine whether or not Austrosaurus is a titanosaur, let alone

1061 a somphospondylan. However, given that the presence of a ventral keel in anterior dorsal

1062 vertebrae is only known in somphospondylan titanosauriforms, we very tentatively suggest

1063 that Austrosaurus might be a member of Somphospondyli. Despite these difficulties, as

1064 elucidated above, the rearticulation of the specimen has facilitated the recognition of several

1065 features of Austrosaurus mckillopi, including one that appears to be autapomorphic (i.e., the

1066 accessory lateral pneumatic foramen on dorsal vertebra I).

1067

1068 Comparison of Austrosaurus mckillopi with other Australian Cretaceous sauropods

1069 The fact that the type specimen of Austrosaurus mckillopi is limited to a posterior

1070 cervical vertebra, the first five dorsal vertebrae, the first five left dorsal ribs and right dorsal

1071 rib IV means that the scope for comparison with Wintonotitan, Diamantinasaurus and

1072 Savannasaurus—the only other named Australian Cretaceous sauropods to date—is limited.

1073 All three of these genera are derived from the Winton Formation, which is at least four

1074 million years younger than the Allaru Mudstone from which Austrosaurus was recovered. On

1075 the basis of this temporal separation alone, it might seem unlikely that Austrosaurus is

1076 congeneric with Diamantinasaurus, Wintonotitan or Savannasaurus; however, this cannot be

1077 ruled out a priori, especially given that several sauropod genera from the Upper Jurassic

1078 Morrison Formation of western North America have stratigraphic ranges that span five

1079 million years or more (Foster 2007).

1080 The few recognisable portions of presacral vertebrae and ribs catalogued as part of the

1081 type specimen of Wintonotitan wattsi (QM F7292) are extremely fragmentary and poorly

1082 preserved, and the only additional specimen referred to Wintonotitan (QM F10916)

1083 comprises four caudal vertebrae (Hocknull et al. 2009, Poropat et al. 2015a). Consequently,

1084 substantive comparisons between Wintonotitan and Austrosaurus cannot be made at this time.

1085 Of the two dorsal vertebrae from the type specimen of Diamantinasaurus matildae

1086 (AODF 603), one (dorsal vertebra B sensu Poropat et al. 2015b) is from the anterior half of

1087 the dorsal series (as interpreted by Poropat et al. 2016, p. 5) based on the position of the

1088 parapophysis (i.e., entirely on the neural arch). The few reasonably complete dorsal vertebrae

1089 preserved in a specimen referred to Diamantinasaurus (AODF 836) appear to have occupied

1090 positions posterior to dorsal vertebra IV (Poropat et al. 2016); thus, they cannot be

1091 substantively compared with Austrosaurus. The preserved dorsal vertebrae of Savannasaurus

1092 elliottorum have been interpreted as dorsals III–X (on the assumption that a total of ten dorsal

1093 vertebrae were present), meaning that the anterior three can be compared with Austrosaurus

1094 (Poropat et al. 2016), although the poor preservation in the latter of dorsal vertebra V

1095 precludes meaningful comparison.

1096 As in Austrosaurus, the centra of dorsal vertebrae III and IV of Savannasaurus, and of

1097 III in Diamantinasaurus, lack ventral ridges (Poropat et al. 2015b, 2016). In the type

1098 specimen of Diamantinasaurus, the ventral surface is concave, both anteroposteriorly and

1099 transversely, and is bounded laterally by ridges (Poropat et al. 2015b); the same is true of

1100 dorsal vertebrae III and IV in Austrosaurus, whereas in Savannasaurus the ventral surfaces of

1101 the dorsal centra are transversely convex and anteroposteriorly concave (Poropat et al. 2016).

1102 In all three taxa, the centra of dorsal vertebrae III and IV are dorsoventrally compressed and

1103 strongly opisthocoelous, and the pneumatic foramina are set within fossae. Comparisons of

1104 the neural arches and laminae systems are limited because of the poor preservation of the

1105 Austrosaurus type series, although some observations can still be made.

1106 The dorsal vertebral lamina systems of Austrosaurus, Diamantinasaurus and

1107 Savannasaurus show several similarities. In all three taxa, dorsal vertebra III appears to

1108 possess two PCPLs, one lower and one upper: the former runs effectively horizontally, dorsal

1109 to the pneumatic foramen, whereas the upper projects anterodorsally. Also projecting

1110 anterodorsally, albeit at a steeper angle than the upper PCPL, is the PCDL. The CDF, which

1111 is bounded by the upper PCPL and the PCDL, is very similar in morphology and position in

1112 both Austrosaurus and Savannasaurus but does not seem to be developed in

1113 Diamantinasaurus.

1114 The congruence between the morphology of the dorsal vertebrae of Austrosaurus,

1115 Diamantinasaurus and Savannasaurus might be indicative of close phylogenetic proximity.

1116 However, it is also possible that these similarities are merely plesiomorphic characters, rather

1117 than shared derived features. None of the features that have been observed in the type series

1118 of Austrosaurus are autapomorphic for either Savannasaurus or Diamantinasaurus, and the

1119 possibility that any two or all three of these taxa are synonymous is remote in light of the

1120 morphological differences observed between Diamantinasaurus and Savannasaurus (Poropat

1121 et al. 2016) and the aforementioned stratigraphic and temporal separation of Austrosaurus

1122 (upper Albian, Allaru Mudstone) from Diamantinasaurus and Savannasaurus (Cenomanian–

1123 lowermost Turonian, Winton Formation). We regard Austrosaurus as a distinct, tentatively

1124 diagnostic taxon, clearly referable to Titanosauriformes. Given that Titanosauriformes by

1125 phylogenetic definition comprises the sister clades Brachiosauridae and Somphospondyli

1126 (Wilson & Sereno 1998), Austrosaurus by extension of this fact has its affinities within one

1127 of these two groups. The hypodigm of Austrosaurus does not present any unambiguous

1128 synapomorphies of the Brachiosauridae. However, absence of evidence alone is not sufficient

1129 grounds to conclusively exclude Austrosaurus from Brachiosauridae, especially given its

1130 incompleteness and the impossibility of verifying characters in materials that are not

1131 preserved. Nonetheless, several aspects of the morphologies we have discussed (i.e., ventral

1132 ridges, patterns of pneumaticity, and dorsal ribs profile) are more consistent with a patchy

1133 distribution among somphospondylan, rather than brachiosaurid, titanosauriform taxa. Thus,

1134 while acknowledging that these aspects of morphology do not represent unambiguous

1135 apomorphies of the clade, Austrosaurus is probably assignable to Somphospondyli

1136 (‘?Somphospondyli incertae sedis’). With future work on character distributions and

1137 discovery of new materials, Austrosaurus might in time even be shown to occupy a position

1138 within Titanosauria

1139

1140 Is the “Hughenden sauropod” cervical vertebra (QM F6142) referable to Austrosaurus?

1141 ===PLEASE INSERT FIGURE 15===

1142 On the basis of our reassessment of the type specimen of Austrosaurus mckillopi,

1143 there is now arguably anatomical overlap between QM F2316, which includes the ultimate

1144 cervical vertebra, and the “Hughenden sauropod” (QM F6142) which solely comprises a

1145 posterior cervical vertebra, possibly the ultimate one. Their spatiotemporal proximity

1146 warrants an assessment of whether or not QM F6142 pertains to Austrosaurus, a suggestion

1147 tentatively made by Molnar (1982a, 1991).

1148 The enigmatic “Hughenden sauropod” is represented only by a single, incomplete

1149 cervical vertebra (QM F6142; Fig. 15). This specimen was collected in 1955 by Jack Tunstall

1150 Woods [then Assistant Curator in Geology at Queensland Museum (Mather 1986)] near

1151 Pelican Bore on Stewart Creek, Dunraven Station, north of Hughenden, Queensland (as

1152 alluded to by Rich & Vickers-Rich 2003, p. 67, and Willis & Thomas 2005, p. 187). As far as

1153 we can ascertain, QM F6142 was first referred to in the literature by Molnar (1980, pp. 132,

1154 136), who identified it as a sauropod from the Albian beds of Queensland and stated that it

1155 was distinct from the sauropods found near Winton. Bartholomai & Molnar (1981, p. 319)

1156 stated that QM F6142 derived from the Wilgunya Subgroup [which includes the Toolebuc

1157 Formation (Vine et al. 1967)], whereas Coombs & Molnar (1981, p. 351) simply noted that it

1158 was from the Lower Cretaceous.

1159 Molnar (1982a, p. 201, 1991, p. 645) provided the first brief description of QM

1160 F6142, raising the possibility that it might represent an additional specimen of Austrosaurus,

1161 but also likening it to Brachiosaurus brancai [now Giraffatitan (Taylor 2009)]. Molnar

1162 (1982a, p. 198, 1991, p. 646) was also the first to illustrate QM F6142, and he reconstructed

1163 the vertebra as being extremely elongate.

1164 Although QM F6142 was listed and briefly alluded to in several checklists of

1165 Australian fossil vertebrates (Molnar 1982b, 1984a, Lees 1986), and likened to

1166 brachiosaurids by Long (1998), it was not discussed by Molnar (2001b) in his otherwise

1167 comprehensive review of Cretaceous sauropod specimens from Queensland. Molnar &

1168 Salisbury (2005) provided a brief description of QM F6142, suggesting that it represented a

1169 brachiosaurid on the basis of several character states listed by Upchurch (1998), Wilson &

1170 Sereno (1998) and Wilson (2002). More recently, Mannion et al. (2013, p. 154) briefly

1171 reassessed QM F6142 and concluded that it was an indeterminate titanosauriform, noting that

1172 no brachiosaurid synapomorphies were observable.

1173

1174 TITANOSAURIFORMES Salgado et al., 1997

1175

1176 Titanosauriformes indet.

1177

1178 Material: QM F6142 (“Hughenden sauropod”), posterior portion of a posterior cervical

1179 vertebra.

1180 Locality: Pelican Bore, Stewart Creek, Dunraven Station, Hughenden, Queensland, Australia.

1181 Collected by Jack Tunstall Woods in 1955.

1182 Horizon: Toolebuc Formation, upper Lower Cretaceous (upper Albian).

1183 Associated vertebrate fauna: Fossil vertebrates recorded from the Toolebuc Formation on

1184 Dunraven Station (as mapped by Vine et al. 1970) include: at least six genera of fish (Lees &

1185 Bartholomai 1987, Bartholomai 2004, 2010a, b, 2012, 2013); turtles (Gaffney 1981),

1186 including Bouliachelys suteri (Kear & Lee 2006); the ichthyosaur Platypterygius australis

1187 (Kear 2001a, 2005, Zammit et al. 2010, Kear & Zammit 2014); an indeterminate

1188 elasmosaurid (Kear 2001b, 2003, Zammit et al. 2008); the pliosaur Kronosaurus

1189 queenslandicus (McHenry 2009); and an ankylosaur [represented by at least two specimens

1190 (AM F35259, AM F119849)] previously assigned to Minmi sp. (Molnar 1996b, Leahey &

1191 Salisbury 2013, Leahey et al. 2015). The only other sauropod specimen known from the

1192 Toolebuc Formation on Dunraven Station is an isolated, amphicoelous caudal centrum (QM

1193 F13712; Molnar & Salisbury 2005).

1194 Description: QM F6142 preserves the posterior cotyle, both postzygapophyses, the partial

1195 neural spine, and portions of several laminae. The posterior cotyle is dorsoventrally

1196 compressed and strongly concave; this latter observation suggests that the centrum was

1197 opisthocoelous, as are all sauropod post-axial cervical vertebrae (McIntosh 1990, Upchurch

1198 et al. 2004). Small sections of the dorsal margin of the cotyle have been broken, revealing

1199 large (35–45 mm long), triangular internal pneumatic coels. The ventral surface of the

1200 centrum is transversely and anteroposteriorly concave, and the preserved portion lacks a

1201 ventral midline ridge. On the ventrolateral surface of the centrum, a horizontal posterior

1202 centroparapophyseal lamina (PCPL) can be observed, strongly suggesting that the

1203 parapophyses were located at the anteroventral corners of the centrum, as in all sauropod

1204 cervical vertebrae. The lateral faces of the centrum are otherwise dominated by posteriorly

1205 acuminate pneumatic fossae; the anterior extent of the pneumatic foramen within each fossa

1206 is obscured on both sides by matrix. On the left side of the vertebra, an accessory fossa (40

1207 mm long) is present near the anteroventral margin of the pneumatic foramen, set within the

1208 pneumatic fossa; this fossa is evidently quite shallow and might simply have been cut off

1209 from the main fossa by an oblique lamina. The posterior centrodiapophyseal laminae

1210 (PCDLs) are well-developed and oriented anterodorsally–posteroventrally. An essentially

1211 vertical anterior centrodiapophyseal lamina (ACDL) can also be observed; the triangular area

1212 ventral to the intersection of these laminae at the diapophysis is filled with matrix but was

1213 evidently a deep centrodiapophyseal fossa (CDF), ventrally bounded by a thin ridge which

1214 separates the CDF from the lateral pneumatic fossa of the centrum. The CDF is best observed

1215 on the left side of the vertebra.

1216 The neural canal is oval in posterior aspect (slightly broader transversely than tall

1217 dorsoventrally), being bounded laterally by the centropostzygapophyseal laminae (CPOLs).

1218 The dorsal margin of the neural canal is obscured by matrix, as are the seemingly shallow

1219 postzygapophyseal centrodiapophyseal fossae (POCDFs). The CPOLs are oriented vertically,

1220 and each CPOL is broad ventrally and narrows dorsally; both are broken at their narrowest

1221 points. The CPOLs deflect slightly laterally at their apices to contact the medial margins of

1222 the postzygapophyses.

1223 The large, flat articular facets of the postzygapophyses face ventrally and somewhat

1224 laterally. The postzygapophyseal facets are ovate, with the lateral margin of each being more

1225 rounded than the medial one. Each postzygapophysis is significantly wider mediolaterally

1226 (164 mm) than long anteroposteriorly (82 mm). The postzygapophyses are attached to the

1227 neural spine via spinopostzygapophyseal laminae (SPOLs). The SPOLs are as broad ventrally

1228 as their respective postzygapophyseal facets and narrow only slightly as they approach the

1229 summit of the neural spine, each being broader than the neural canal where they converge

1230 with the neural spine; consequently, the SPOLs in QM F6142 would be better referred to as

1231 buttresses rather than laminae. Epipophyses are not present on the postzygapophyses or on

1232 the SPOLs. Towards the summit of the neural spine, both SPOLs split into two branches,

1233 defining spinopostzygapophyseal lamina fossae (SPOL-Fs), which are filled with matrix; the

1234 depth of these cannot be ascertained. The SPOLs and the dorsal margin of the neural canal

1235 define the boundaries of the spinopostzygapophyseal fossa (SPOF), which is not deeply

1236 invaginated on the posterior surface of the neural spine. No trace of a postspinal lamina

1237 (POSL) is preserved within the SPOF. At the anterior margin of the right postzygapophysis, a

1238 posterodorsally–anteroventrally extending postzygodiapophyseal lamina (PODL) is

1239 preserved. On its anterolateral surface, the right SPOL defines a narrow ridge that represents

1240 the posterior margin of an apparently short, rounded spinodiapophyseal fossa (SDF), which

1241 was bounded ventrally by the PODL. The anterior extent of this SDF is unknown because of

1242 the incompleteness of the specimen. As far as we can determine, the neural spine of QM

1243 F6142 was not bifid. If it were, the notch would have been small and restricted to the summit

1244 of the spine.

1245 Most of the anterior portion of QM F6142 is missing, and this surface has also only

1246 been incompletely prepared; nevertheless, some observations can be made. The junction of

1247 the truncated PCPLs and the truncated combined base of the ACDLs and

1248 centroprezygapophyseal laminae (CPRLs) defines an X-shape in anterior view. As far as they

1249 are preserved, the bases of the CPRLs are broad mediolaterally. On the lateral surface,

1250 shallow triangular fossae appear to be defined on both sides by the ACDLs (posteroventrally)

1251 and CPRLs (anteroventrally); prezygodiapophyseal laminae (PRDLs) are presumed to have

1252 been present and would have formed the dorsal margins of these fossae (the base of the right

1253 PRDL has been tentatively identified). Consequently, we interpret these fossae as

1254 prezygapophyseal centrodiapophyseal fossae (PRCDFs), assuming no additional laminae

1255 were present that would warrant alternative identifications (as per Wilson et al. 2011b). On

1256 the right side, the CDF and PRCDF are similar in size. The exposed neural canal opening is

1257 oval (being wider transversely than tall dorsoventrally).

1258 Comparisons: The SPOL-Fs of QM F6142 appear to be autapomorphic. Although SPOL-Fs

1259 have been reported in other sauropods (Wilson et al. 2011b, Ibiricu et al. 2013, Mannion &

1260 Barrett 2013), these are morphologically divergent from those of QM F6142 and are present

1261 only in the posterior dorsal vertebrae.

1262 The widely-spaced, ventrally-facing postzygapophyses indicate that QM F6142 is a

1263 posterior cervical vertebra—potentially the posteriormost. Consequently, it is possible that

1264 QM F6142 overlaps anatomically with the anteriormost vertebra preserved in the

1265 Austrosaurus mckillopi type specimen. The accessory fossa within the lateral pneumatic fossa

1266 observed on the left side of QM F6142 does not seem to be homologous with the accessory

1267 foramen identified in dorsal vertebra I of Austrosaurus—that in QM F6142 was situated

1268 posterior to the parapophysis, not anterior to it as in Austrosaurus. The posterior cotyles of

1269 both QM F2316 and QM F6142 are dorsoventrally compressed, and the internal texture of

1270 both specimens is camellate; these observations are, however, not sufficient to unequivocally

1271 refer QM F6142 to Austrosaurus. Future discoveries in the Toolebuc Formation and the

1272 Allaru Mudstone might shed further light on the diversity of sauropods in the latest Early

1273 Cretaceous of northeast Australia, and on whether or not QM F6142 is referable to

1274 Austrosaurus. If a series of presacral vertebrae were discovered, wherein SPOL-Fs were

1275 present in the posterior cervicals and accessory lateral pneumatic foramina were found

1276 anterior to the parapophyses in dorsal vertebra I, then referral of QM F6142 to Austrosaurus

1277 mckillopi could be confirmed.

1278 The pneumatic fossae on both sides of QM F6142 appear to be almost completely

1279 preserved. This, coupled with the observation of the posteriormost section of the PCPL on the

1280 left lateral surface, suggests that QM F6142 was not as elongate as restored by Molnar

1281 (1982a, 1991). Furthermore, the posterior cotyle of the “Hughenden sauropod” (Table 3) is

1282 actually smaller than the posterior cotyle of the posterior cervical of Austrosaurus mckillopi

1283 (Table 1). Size estimates of the sauropod from which QM F6142 derived [20 m according to

1284 Molnar (1982a, 1991)] are, therefore, probably excessive.

1285

1286 Possible biogeographic links between Austrosaurus mckillopi and South American Early and

1287 earliest Late Cretaceous titanosauriforms

1288 Poropat et al. (2016) suggested that some sauropod clades, specifically

1289 titanosauriforms, might have taken advantage of late Albian–Turonian warming to migrate

1290 between South America and Australia via Antarctica, with Austrosaurus possibly

1291 representing one such migrant. Testing this hypothesis is, however, difficult because of the

1292 relative dearth of Early Cretaceous titanosauriform body fossils from South America, and the

1293 complete lack of such fossils from Lower Cretaceous Antarctic and pre-Albian Australian

1294 strata. The oldest known titanosauriform specimens from South America are of Hauterivian–

1295 Barremian age, and only seven pre-Cenomanian deposits have yielded titanosauriforms to

1296 date (de Jesus Faria et al. 2015). The specimens from these units, along with those from the

1297 lower Cenomanian Candeleros Formation (Rio Limay Subgroup, Neuquén Group), are

1298 briefly discussed here to provide somewhat limited phylogenetic and biogeographic context

1299 for Austrosaurus and other mid-Cretaceous sauropods from Australia.

1300 Two presacral vertebral centra, a tibia and an indeterminate limb bone from the

1301 Hauterivian–Barremian Puesto La Paloma Member of the Cerro Barcino Formation of

1302 Chubut (Rauhut et al. 2003) constitute the oldest reported body fossils of titanosauriforms

1303 from Argentina. Although these specimens were originally described as pertaining to

1304 titanosaurs, the justification for this referral is weak. The morphology of the lateral pneumatic

1305 foramen of the centrum (elongate and eye-shaped) was the only character used to support

1306 titanosaur affinities for the vertebrae (following Salgado et al. 1997); however, the actual

1307 shape of the foramina in these vertebrae was not observed—it was inferred on the basis of

1308 their length (Rauhut et al. 2003). Furthermore, some non-titanosaurian titanosauriforms also

1309 possess this feature [e.g. Chubutisaurus (Carballido et al. 2011a), Europasaurus (Carballido

1310 & Sander 2014), Sauroposeidon (D'Emic & Foreman 2012)], so it is more appropriate to

1311 consider these Argentinean specimens as indeterminate titanosauriforms.

1312 A right fibula and a partial skeleton from the Rio Piranhas Formation (Hauterivian–

1313 Barremian) of Paraíba constitute the oldest reported titanosauriform body fossils from Brazil

1314 (Ghilardi et al. 2016). The isolated fibula was described as a titanosaur; however, the

1315 evidence for this is weak. Ghilardi et al. (2016) suggested that the distal end being triangular,

1316 and the overall shape of the fibula being sigmoidal, were sufficient grounds for referral of this

1317 element to Titanosauria. However, sigmoidal fibulae were found to be synapomorphic for

1318 Somphospondyli/Titanosauria by Mannion et al. (2013) in their LCDM analysis [wherein

1319 Somphospondyli and Titanosauria were interchangeable because Andesaurus delgadoi (the

1320 most basal titanosaur by definition) clustered with a group of sauropods otherwise commonly

1321 resolved as non-titanosaurian somphospondylans], whereas D'Emic (2012) identified

1322 sigmoidal fibulae as synapomorphic for an unnamed clade comprising Tastavinsaurus,

1323 Euhelopodidae, Chubutisaurus and Titanosauria, thereby encompassing most of

1324 Somphospondyli. It is unlikely that the fibula described by Ghilardi et al. (2016) will ever be

1325 referred unequivocally to Titanosauria; it is, however, possible to tentatively refer it to

1326 Somphospondyli.

1327 Also from the Rio Piranhas Formation is the type specimen of Triunfosaurus

1328 leonardii, which comprises a right ischium, three caudal vertebrae, three chevrons and three

1329 isolated neural spines (Carvalho et al. 2017). Triunfosaurus was interpreted as a titanosaur on

1330 morphological grounds, and was resolved as such in a phylogenetic analysis [based on the

1331 data matrix of Carballido & Sander (2014)] by Carvalho et al. (2017). However, the case for

1332 Triunfosaurus as a titanosaur is weak. The pubic articulation of the right ischium is longer

1333 than the anteroposterior length of the iliac peduncle, a feature used by Carvalho et al. (2017)

1334 to justify referral to Camarasauromorpha [following Salgado et al. (1997)]. The neural arches

1335 of the caudal vertebrae are situated upon the anterior halves of the centra, a feature used by

1336 Carvalho et al. (2017) to refer Triunfosaurus to a clade comprising Europasaurus holgeri and

1337 all more derived camarasauromorphs [following Carballido & Sander (2014)]. The proximal

1338 articular surfaces of at least one of the chevrons were each divided into two discrete surfaces

1339 by a furrow, and Carvalho et al. (2017) used this to support the notion that Triunfosaurus was

1340 a titanosaur, citing Mannion & Calvo (2011) to support this. However, the feature Mannion

1341 & Calvo (2011) described as being present in Andesaurus was a strong convexity dividing the

1342 two surfaces, as also observed in Tastavinsaurus (Canudo et al. 2008) and the non-

1343 neosauropod Cetiosaurus (Upchurch & Martin 2002). The furrow morphology is present in

1344 some titanosaurs [e.g. Aeolosaurus (Powell 2003, Santucci & Arruda-Campos 2011) and

1345 Epachthosaurus (Poropat et al. 2016)]; however, it has also been identified in the non-

1346 titanosaurian somphospondylans Phuwiangosaurus and Tangvayosaurus (D'Emic 2012).

1347 Consequently, this feature supports the inclusion of Triunfosaurus within Somphospondyli

1348 but does not allow unequivocal referral to Titanosauria. Carvalho et al. (2017) also suggested

1349 that the caudal prezygapophyses, which project anteriorly, are reminiscent of Aeolosaurini.

1350 The caudal centra were interpreted as opisthoplatyan [the anterior articular surfaces could not

1351 be observed, according to Carvalho et al. (2017)], separating Triunfosaurus from the majority

1352 of titanosaurs (Upchurch et al. 2004), and the dorsoventral height of the haemal canal was

1353 found to be less than 50% the overall length of the chevron, in contrast to Titanosauria as

1354 characterised by Wilson (2002); however, the latter feature has been shown to much more

1355 variable (Mannion et al. 2013). There are a number of features of these caudal vertebrae

1356 which were not mentioned by Carvalho et al. (2017), which are quite notable: 1) the caudal

1357 vertebrae have prominent transverse processes, connected to the prezygapophyses by

1358 pronounced PRDLs, which terminate in deep diapophyses (for middle caudal vertebrae); 2)

1359 the prezygapophyses are connected to the neural spine by well-developed SPRLs [which are

1360 omitted from the schematic provided by Carvalho et al. (2017, fig. 4)]; 3) the ventrolateral

1361 surface of the centrum appears to be deeply excavated (Carvalho et al. 2017, fig. 4); and 4)

1362 the anterior margin of the postzygapophyseal facet is situated in line with the midlength of

1363 the centrum. All of these features are unusual in middle caudal vertebrae of sauropods

1364 generally. Furthermore, the presence of such a prominent transverse process implies that

1365 these vertebrae were situated more anteriorly within the tail than postulated by Carvalho et al.

1366 (2017, fig. 7). If we presume that this was so, then these caudal vertebrae are perhaps too

1367 small to be associated with the ischium. In sum, Triunfosaurus is a problematic taxon, albeit

1368 one which appears to be referable to Somphospondyli.

1369 Several teeth from the Barremian–lower Aptian La Amarga Formation of Neuquén,

1370 Argentina have been assigned to Titanosauria (Apesteguía 2007). However, this

1371 interpretation has been questioned (Zaher et al. 2011), with some workers explicitly

1372 removing them from Titanosauria (D'Emic 2012) and others hesitating to classify them

1373 beyond Titanosauriformes (Gallina 2016). Amargatitanis macni, a sauropod also derived

1374 from the La Amarga Formation, was originally described as a titanosaur (Apesteguía 2007);

1375 however, a full reappraisal of the type specimen has revealed it to be a chimaera, as suggested

1376 by D'Emic (2012), with the majority of the remains actually pertaining to a dicraeosaurid

1377 diplodocoid (Gallina 2016).

1378 The titanosauriform Padillasaurus leivaensis was erected on the basis of fourteen

1379 vertebrae (two dorsals, four sacrals and eight caudals) from the Barremian–Aptian Paja

1380 Formation of Colombia, and referred to Brachiosauridae (Carballido et al. 2015). The internal

1381 morphology of the vertebrae, which have coels both large (camerae) and small (camellae) and

1382 are therefore semicamellate, lends support to this interpretation. However, another feature

1383 utilised by these authors to support its referral to Brachiosauridae, i.e. the blind lateral fossa

1384 in the caudal centra, is also present in the titanosaur Savannasaurus (Poropat et al. 2016), and

1385 a recent analysis has recovered Padillasaurus as a non-titanosaurian somphospondylan

1386 (Mannion et al. 2017).

1387 Tapuiasaurus macedoi is represented by a complete skull and partial skeleton from

1388 the Aptian Quiricó Formation of Minas Gerais, Brazil (Zaher et al. 2011, Wilson et al. 2016).

1389 Although the postcranial skeleton of this taxon has only been perfunctorily described to date,

1390 it has been made clear that the presacral vertebrae have camellate internal texture, the

1391 proximal ends of the dorsal ribs are pneumatised, and the distal ends of the anterior dorsal

1392 ribs are plank-like in cross-section (Zaher et al. 2011, Wilson et al. 2016). Tapuiasaurus has

1393 consistently been recovered within Titanosauria in phylogenetic analyses. However, although

1394 virtually every phylogenetic analysis in which it has been included has resolved it within

1395 Lithostrotia (Zaher et al. 2011, Carballido & Sander 2014, Gorscak et al. 2014, Lacovara et

1396 al. 2014, Carballido et al. 2015, Poropat et al. 2015b, Díez Díaz et al. 2016, González Riga et

1397 al. 2016, Gorscak & O'Connor 2016, Martínez et al. 2016, Poropat et al. 2016, Filippini et al.

1398 2017, Tykoski & Fiorillo 2017), more recent research on the cranial remains suggests that

1399 Tapuiasaurus might occupy a basal position within Titanosauria (Wilson et al. 2016), outside

1400 Lithostrotia [or within Somphospondyli but just outside Titanosauria—the omission of

1401 Andesaurus from the phylogenetic analyses of Wilson et al. (2016) means that the placement

1402 of the node Titanosauria is subjective].

1403 The upper Aptian–lower Albian Lohan Cura Formation of Neuquén, Argentina, has

1404 produced abundant sauropod remains, including Agustinia ligabuei (Bonaparte 1999, Salgado

1405 & Coria 2005, Salgado & Bonaparte 2007) and Ligabuesaurus leanzai (Bonaparte et al.

1406 2006, Martinelli et al. 2007). The type specimen of Agustinia is incomplete, poorly preserved

1407 and difficult to interpret; consequently, it has variously been regarded as an indeterminate

1408 neosauropod (D'Emic et al. 2009), a nomen dubium (D'Emic 2012), and an indeterminate

1409 somphospondylan (Mannion et al. 2013). D'Emic et al. (2009) questioned the interpretation

1410 of the osteoderms, suggesting that they might perhaps be hypertrophied ossifications,

1411 whereas Mannion et al. (2013) suggested that the elements represented dorsal ribs and pelvic

1412 girdle elements. A recent histological study of the type specimen (Bellardini & Cerda 2017)

1413 has concluded that Mannion et al. (2013) were correct: the supposed armour of Agustinia

1414 comprises misinterpreted ribs and pelvic girdle elements, thereby undermining the evidence

1415 for its inclusion in Lithostrotia.

1416 Ligabuesaurus is known from more complete and better preserved material than

1417 Agustinia and has generally been resolved as a basal somphospondylan (Bonaparte et al.

1418 2006, D'Emic 2012 and references therein, Mannion et al. 2013, Carballido & Sander 2014,

1419 Poropat et al. 2015b, Díez Díaz et al. 2016, González Riga et al. 2016, Poropat et al. 2016),

1420 although a small number of analyses have placed it within Titanosauria (Carballido et al.

1421 2015, Gorscak & O'Connor 2016). The type specimen of Ligabuesaurus includes an anterior

1422 dorsal vertebra; unfortunately, the ventral surface of the centrum was not described by

1423 Bonaparte et al. (2006), and the presence or absence of a ventral ridge could not be

1424 determined even through firsthand personal observation of the specimen (by PDM).

1425 The lower/middle Albian Itapecuru Group of Maranhão, Brazil has produced very

1426 fragmentary sauropod remains that have been assigned to Titanosauria (Castro et al. 2007).

1427 The dorsal vertebrae were classified as such on the basis of their internal texture, which is

1428 semicamellate, whereas the amphicoelous caudal centrum was assigned to Titanosauria

1429 because the neural arch is situated anteriorly. Both of these features are now known to occur

1430 more widely in Titanosauriformes, suggesting that these specimens cannot be referred

1431 unequivocally to Titanosauria.

1432 The lower Cenomanian Candeleros Formation of Neuquén, Argentina has produced

1433 Andesaurus delgadoi (Calvo & Bonaparte 1991, Mannion & Calvo 2011), the most basal

1434 titanosaur by definition (Wilson & Upchurch 2003). Epachthosaurus sciuttoi, a titanosaur

1435 best known from the Lower Member of the Bajo Barreal Formation (Cenomanian–Turonian)

1436 of Chubut, Argentina (Powell 1990, 2003, Martínez et al. 2004), has also been reported from

1437 this unit (Salgado & Coria 2005, Salgado & Bonaparte 2007); however, the specimens that

1438 might indicate its presence in the Candeleros Formation have never been described. An

1439 indeterminate titanosaur (MUCPv 271; now stored at MMCH), initially reported by Calvo

1440 (1999) as an additional specimen of Andesaurus, is represented by a partial pelvis and several

1441 caudal vertebrae. On the basis of the morphology of the pubis, Mannion & Calvo (2011)

1442 assigned MUCPv 271 to ‘Titanosauriformes indet.’—the caudal vertebrae previously

1443 reported could not be located. Another specimen once referred to Andesaurus, comprising a

1444 series of caudal vertebrae and associated chevrons (MMCH-Pv 47), was regarded as an

1445 indeterminate titanosaur by Otero et al. (2011).

1446 At this stage, meaningful comparison between Austrosaurus and the majority of the

1447 known South American Early Cretaceous titanosauriform specimens is not possible because

1448 few of the specimens overlap anatomically. Despite this, our brief summary of the Early

1449 Cretaceous South American titanosauriform body fossil record demonstrates that multiple

1450 titanosauriform taxa existed in South America prior to the end of the Albian. Any or all of the

1451 clades to which these taxa pertain might have been able to take advantage of high latitude

1452 dispersal routes to Australia via Antarctica when conditions were favourable. However,

1453 because the precise position of Austrosaurus within Titanosauriformes is unknown, and is

1454 difficult to resolve on the basis of the type material alone, we will be forced to rely upon

1455 future discoveries to precisely determine the palaeobiogeographic significance of

1456 Austrosaurus.

1457

1458 Conclusion

1459 The sauropod taxon Austrosaurus mckillopi is of historical significance to Australian

1460 palaeontology as the first Cretaceous dinosaur recognised in Queensland, and the first

1461 Cretaceous sauropod ever reported from the entire continent. The augmentation, articulation

1462 and description of the type material have helped to shed light on the phylogenetic position of

1463 Austrosaurus, unequivocally placing it within Titanosauriformes, and probably as a member

1464 of Somphospondyli. The identification of an autapomorphic auxiliary pneumatic foramen in

1465 dorsal vertebra I means that the referral of other sauropod specimens to Austrosaurus should

1466 be possible in the future, although this feature is not presently observable in any other

1467 Australian sauropod specimen. The morphological congruence of the posteriormost cervical

1468 vertebra of Austrosaurus with QM F6142 (the “Hughenden sauropod”) might represent

1469 grounds for the referral of the latter to the former, although this cannot be demonstrated

1470 unequivocally. Lastly, despite its fragmentary nature, Austrosaurus appears to share several

1471 features with the type specimens of both Diamantinasaurus and Savannasaurus, possibly

1472 indicating a close phylogenetic relationship.

1473 The fragmentary nature of the type series of Austrosaurus has impeded, and will

1474 continue to restrict, efforts to precisely resolve its phylogenetic position within

1475 Titanosauriformes. Consequently, the palaeobiogeographic significance of Austrosaurus is

1476 poorly understood, a situation worsened by the relative rarity of Early Cretaceous

1477 titanosauriforms in South America and the lack of such in Antarctica. Nevertheless, the

1478 presence of numerous titanosauriform lineages in the Early Cretaceous of South America

1479 provides some context for Australian Early Cretaceous titanosauriforms like Austrosaurus,

1480 and also for the mid-Cretaceous Winton Formation fauna, which appears to have been

1481 dominated by titanosaurs with amphicoelous (rather than procoelous) caudal vertebrae.

1482

1483 Acknowledgments

1484 The authors would like to thank: Eric & Lynne Slacksmith (Clutha Station) for allowing us

1485 access to the site; John Wharton (Richmond Shire Council) for relocating the site; Gary and

1486 Barb Flewelling, George Sinapius (AAOD), Kathrine Thompson (AAOD), Dennis Clancy,

1487 Alan & Lyn Scrymgeour, Mal & Jane Garden, John & Carry Jenkins, and Annie Just for their

1488 assistance in excavating Austrosaurus; Jose Carlos Mendezona, Tyrell Watson, and Dennis

1489 Friend (all Richmond Shire Council) for operating the earth-moving equipment; Richmond

1490 Shire Council for loaning the earth-moving equipment; H. B. Wade’s sons Peter and Richard

1491 Wade, and Dr M. J. McKillop’s daughters Elizabeth Cleary and Kathryn Evans for provision

1492 of images and information; David & Judy Elliott (AAOD) for provision of information,

1493 contacts and support; Tony Thulborn and Ralph Molnar for personal communications related

1494 to their expeditions to the Austrosaurus site; Felicity Tomlinson and Tanya Mellar (both

1495 AAOD) for providing transport to and from Richmond for SFP; Kristen Spring and Andrew

1496 Rozefelds (Queensland Museum) and Patrick Smith (Kronosaurus Korner) for access to

1497 specimens in their care; Trish Sloan (AAOD) for providing the photographs of the

1498 Diamantinasaurus ribs; Nicole Newman, Queensland X-Ray, Greenslopes Private Hospital,

1499 Rochelle Lawrence and Kristen Spring (both QM) for CT scanning the Austrosaurus

1500 mckillopi type series; and Phil Bell and Ignacio Cerda for their reviews of the manuscript.

1501 PU’s visit to Australia to examine sauropod specimens was supported by Leverhulme Trust

1502 Research Grant RPG-129.

1503

1504

1505 References

1506

1507 ADAMS, T.L. 2009. Deposition and taphonomy of the Hound Island Late Triassic vertebrate

1508 fauna: fossil preservation within subaqueous gravity flows. PALAIOS 24, 603–615.

1509 AGNOLIN, F.L., EZCURRA, M.D., PAIS, D.F. & SALISBURY, S.W. 2010. A reappraisal of

1510 the Cretaceous non-avian dinosaur faunas from Australia and New Zealand: evidence

1511 for their Gondwanan affinities. Journal of Systematic Palaeontology 8, 257–300.

1512 ALLISON, P.A. & BRIGGS, D.E.G., 1991. Taphonomy of nonmineralized tissues. In

1513 Taphonomy: Releasing the Data Locked in the Fossil Record. ALLISON, P.A. &

1514 BRIGGS, D.E.G., eds, Plenum Press, New York, 25–70.

1515 ANDERSON, G.S. & BELL, L.S. 2014. Deep coastal marine taphonomy: investigation into

1516 carcass decomposition in the Saanich Inlet, British Columbia using a baited camera.

1517 PLoS ONE 9, e110710.

1518 ANONYMOUS, 1933a. Fossil of new type of giant dinosaur discovered in Queensland. The

1519 Brisbane Courier March 15th 1933, 14.

1520 ANONYMOUS, 1933b. A gigantic dinosaur fossil. An important discovery at Clutha. The

1521 Brisbane Courier March 15th 1933, 11.

1522 APESTEGUÍA, S. 2007. The sauropod diversity of the La Amarga Formation (Barremian),

1523 Neuquén (Argentina). Gondwana Research 12, 533–546.

1524 BANDEIRA, K.L.N., MEDEIROS SIMBRAS, F., BATISTA MACHADO, E., CAMPOS,

1525 D.D.A., OLIVEIRA, G.R. & KELLNER, A.W.A. 2016. A new giant Titanosauria

1526 (Dinosauria: Sauropoda) from the Late Cretaceous Bauru Group, Brazil. PLoS ONE

1527 11, e0163373.

1528 BARCO, J.L., CANUDO, J.I. & CUENCA-BESCÓS, G. 2006. Descripción de las vértebras

1529 cervicales de Galvesaurus herreroi Barco, Canudo, Cuenca-Bescós & Ruiz-Omeñaca,

1530 2005 (Dinosauria, Sauropoda) del tránsito Jurásico–Cretácico en Galve (Teruel,

1531 España). Revista Española de Paleontología 21, 189–205.

1532 BARCO RODRIGUEZ, J.L., 2009. Sistemática e implicaciones filogenéticas y

1533 paleobiogeográficas del saurópodo Galvesaurus herreroi (Formación Villar del

1534 Arzobispo, Galve, España). Ph.D. thesis. Universidade de Zaragoza. 399 pp.

1535 BARDACK, D. 1962. Taxonomic status and geological position of the Cretaceous fish

1536 Ichthyodectes marathonensis. Australian Journal of Science 24, 387–388.

1537 BARRETT, P.M. & UPCHURCH, P., 2005. Sauropodomorph diversity through time:

1538 paleoecological and macroevolutionary implications. In The Sauropods: Evolution

1539 and Paleobiology. CURRY ROGERS, K.A. & WILSON, J.A., eds, University of

1540 California Press, Berkeley, 125–156.

1541 BARTHOLOMAI, A., 1969. The Lower Cretaceous elopoid fish Pachyrhizodus

1542 marathonensis (Etheridge Jnr.). In Stratigraphy and Palaeontology: Essays in Honour

1543 of Dorothy Hill. CAMPBELL, K.S.W., ed, Australian National University Press,

1544 Canberra, 249–263.

1545 BARTHOLOMAI, A. 2004. The large aspidorhynchid fish Richmondichthys sweeti

1546 (Etheridge Jr and Smith Woodward, 1891) from Albian marine deposits of

1547 Queensland, Australia. Memoirs of the Queensland Museum 49, 521–536.

1548 BARTHOLOMAI, A. 2008. Lower Cretaceous chimaeroids (Chondrichthyes: Holocephali)

1549 from the Great Artesian Basin, Australia. Memoirs of the Queensland Museum 52,

1550 49–56.

1551 BARTHOLOMAI, A. 2010A. A new Albian teleost, Euroka dunravenensis gen. et sp. nov.

1552 and a new family, Eurokidae, from the Eromanga Basin of Queensland. Memoirs of

1553 the Queensland Museum 55, 69–85.

1554 BARTHOLOMAI, A. 2010B. Revision of Flindersichthys denmeadi Longman 1932, a

1555 marine teleost from the Lower Cretaceous of the Great Artesian Basin, Queensland.

1556 Memoirs of the Queensland Museum 55, 43–68.

1557 BARTHOLOMAI, A. 2012. The pachyrhizodontid teleosts from the marine Lower

1558 Cretaceous (latest mid to late-Albian) sediments of the Eromanga Basin, Queensland,

1559 Australia. Memoirs of the Queensland Museum 56, 119–147.

1560 BARTHOLOMAI, A. 2013. New teleosts (Elopomorpha: Albuliformes) from the Lower

1561 Cretaceous (Late Albian) of the Eromanga Basin, Queensland, Australia. Memoirs of

1562 the Queensland Museum 58, 73–94.

1563 BARTHOLOMAI, A. & MOLNAR, R.E. 1981. Muttaburrasaurus, a new iguanodontid

1564 (Ornithischia: Ornithopoda) dinosaur from the Lower Cretaceous of Queensland.

1565 Memoirs of the Queensland Museum 20, 319–349.

1566 BEARDMORE, S.R., ORR, P.J., MANZOCCHI, T. & FURRER, H. 2012. Float or sink:

1567 modelling the taphonomic pathway of marine crocodiles (Mesoeucrocodylia,

1568 Thalattosuchia) during the death–burial interval. Palaeobiodiversity and

1569 Palaeoenvironments 92, 83–98.

1570 BELLARDINI, F. & CERDA, I.A. 2017. Bone histology sheds light on the nature of the

1571 “dermal armor” of the enigmatic sauropod dinosaur Agustinia ligabuei Bonaparte,

1572 1999. The Science of Nature 104.

1573 BONAPARTE, J.F. 1986. Les Dinosaures (Carnosaures, Allosauridés, Sauropodes,

1574 Cétiosauridés) du Jurassique Moyen de Cerro Cóndor (Chubut, Argentine) (2e partie

1575 et fin). Annales de Paléontologie (Vert.-Invert.) 72, 325–386.

1576 BONAPARTE, J.F., 1999. An armoured sauropod from the Aptian of northern Patagonia,

1577 Argentina. In Proceedings of the Second Gondwanan Dinosaur Symposium: National

1578 Science Museum Monograph, 15. TOMIDA, Y., RICH, T.H. & VICKERS-RICH, P.,

1579 eds, National Science Museum, Tokyo, 1–12.

1580 BONAPARTE, J.F. & CORIA, R.A. 1993. Un nuevo y gigantesco saurópodo titanosaurio de

1581 la Formación Rio Limay (Albiano-Cenomaniano) de la Provincia del Neuquén,

1582 Argentina. Ameghiniana 30, 271–282.

1583 BONAPARTE, J.F., GONZÁLEZ RIGA, B.J. & APESTEGUÍA, S. 2006. Ligabuesaurus

1584 leanzai gen. et sp. nov. (Dinosauria, Sauropoda), a new titanosaur from the Lohan

1585 Cura Formation (Aptian, Lower Cretaceous) of Neuquén, Patagonia, Argentina.

1586 Cretaceous Research 27, 364–376.

1587 BORSUK-BIAŁYNICKA, M. 1977. A new camarasaurid sauropod Opisthocoelicaudia

1588 skarzynskii gen. n., sp. n. from the Upper Cretaceous of Mongolia. Palaeontologia

1589 Polonica 37, 5–64.

1590 BRYAN, S.E., COOK, A.G., ALLEN, C.M., SIEGEL, C., PURDY, D.J., GREENTREE, J.S.

1591 & UYSAL, I.T. 2012. Early–mid Cretaceous tectonic evolution of eastern Gondwana:

1592 from silicic LIP magmatism to continental rupture. Episodes 35, 142–152.

1593 BUFFETAUT, E. 1994. The significance of dinosaur remains in marine sediments: an

1594 investigation based on the French record. Berliner Geowissenschaftliche

1595 Abhandlungen E 13, 125–133.

1596 BURGER, D., 1986. Palynology, cyclic sedimentation and palaeoenvironments in the Late

1597 Mesozoic of the Eromanga Basin. In Contributions to the Geology and Hydrocarbon

1598 Potential of the Eromanga Basin: Geological Society of Australia Special Publication

1599 12. GRAVESTOCK, D.I., MOORE, P.S. & PITT, G.M., eds, Geological Society of

1600 Australia, Sydney 53–70.

1601 CALVO, J.O., 1999. Dinosaurs and other vertebrates of the Lake Ezequiel Ramos Mexia

1602 Area, Neuquén - Patagonia, Argentina. In Proceedings of the Second Gondwanan

1603 Dinosaur Symposium. National Science Museum Monograph, 15. TOMIDA, Y.,

1604 RICH, T.H. & VICKERS-RICH, P., eds, National Science Museum, Tokyo, 13–45.

1605 CALVO, J.O. & BONAPARTE, J.F. 1991. Andesaurus delgadoi gen. et sp. nov. (Saurischia–

1606 Sauropoda), dinosaurio Titanosauridae de la Formación Río Limay (Albiano–

1607 Cenomaniano), Neuquén, Argentina. Ameghiniana 28, 303–310.

1608 CALVO, J.O., PORFIRI, J.D., GONZÁLEZ RIGA, B.J. & KELLNER, A.W.A. 2007.

1609 Anatomy of Futalognkosaurus dukei Calvo, Porfiri, González Riga & Kellner, 2007

1610 (Dinosauria, Titanosauridae) from the Neuquén Group (Late Cretaceous), Patagonia,

1611 Argentina. Arquivos do Museu Nacional, Rio de Janeiro 65, 511–526.

1612 CAMPOS, D.D.A., KELLNER, A.W.A., BERTINI, R.J. & SANTUCCI, R.M. 2005. On a

1613 titanosaurid (Dinosauria, Sauropoda) vertebral column from the Bauru Group, Late

1614 Cretaceous of Brazil. Arquivos do Museu Nacional, Rio de Janeiro 63, 565–593.

1615 CANUDO, J.I., ROYO-TORRES, R. & CUENCA-BESCÓS, G. 2008. A new sauropod:

1616 Tastavinsaurus sanzi gen. et sp. nov. from the Early Cretaceous (Aptian) of Spain.

1617 Journal of Vertebrate Paleontology 28, 712–731.

1618 CARBALLIDO, J.L., POL, D., CERDA, I. & SALGADO, L. 2011A. The osteology of

1619 Chubutisaurus insignis Del Corro, 1975 (Dinosauria: Neosauropoda) from the

1620 ‘Middle’ Cretaceous of Central Patagonia, Argentina. Journal of Vertebrate

1621 Paleontology 31, 93–110.

1622 CARBALLIDO, J.L., POL, D., PARRA RUGE, M.L., PADILLA BERNAL, S., PÁRAMO-

1623 FONSECA, M.E. & ETAYO-SERNA, F. 2015. A new Early Cretaceous

1624 brachiosaurid (Dinosauria, Neosauropoda) from northwestern Gondwana (Villa de

1625 Leiva, Colombia). Journal of Vertebrate Paleontology 35, e980505.

1626 CARBALLIDO, J.L., RAUHUT, O.W.M., POL, D. & SALGADO, L. 2011B. Osteology and

1627 phylogenetic relationships of Tehuelchesaurus benitezii (Dinosauria, Sauropoda) from

1628 the Upper Jurassic of Patagonia. Zoological Journal of the Linnean Society 163, 605–

1629 662.

1630 CARBALLIDO, J.L., SALGADO, L., POL, D., CANUDO, J.I. & GARRIDO, A. 2012. A

1631 new basal rebbachisaurid (Sauropoda, Diplodocoidea) from the Early Cretaceous of

1632 the Neuquén Basin; evolution and biogeography of the group. Historical Biology 24,

1633 631–654.

1634 CARBALLIDO, J.L. & SANDER, P.M. 2014. Postcranial axial skeleton of Europasaurus

1635 holgeri, (Dinosauria, Sauropoda) from the Upper Jurassic of Germany: implications

1636 for sauropod ontogeny and phylogenetic relationships of basal Macronaria. Journal of

1637 Systematic Palaeontology 12, 335–387.

1638 CARVALHO, I.D.S., SALGADO, L., LINDOSO, R.M., ARAÚJO-JÚNIOR, H.I.D.,

1639 NOGUEIRA, F.C.C. & SOARES, J.A. 2017. A new basal titanosaur (Dinosauria,

1640 Sauropoda) from the Lower Cretaceous of Brazil. Journal of South American Earth

1641 Sciences 75, 74–84.

1642 CASTRO, D.F., BERTINI, R.J., SANTUCCI, R.M. & MEDEIROS, M.A. 2007. Sauropods of

1643 the Itapecuru Group (Lower / Middle Albian), São Luís-Grajaú Basin, Maranhão

1644 State, Brazil. Revista Brasileira de Paleontologia 10, 195–200.

1645 CERDA, I.A., SALGADO, L. & POWELL, J.E. 2012. Extreme postcranial pneumaticity in

1646 sauropod dinosaurs from South America. Paläontologische Zeitschrift 86, 441–449.

1647 COOK, A.G. 2008. Asteriacites in the Lower Cretaceous of Queensland. Memoirs of the

1648 Queensland Museum 52, 88.

1649 COOK, A.G. 2012. Cretaceous faunas and events, northern Eromanga Basin, Queensland.

1650 Episodes 35, 153–159.

1651 COOMBS, W.P., JR. & MOLNAR, R.E. 1981. Sauropoda (Reptilia, Saurischia) from the

1652 Cretaceous of Queensland. Memoirs of the Queensland Museum 20, 351–373.

1653 CORIA, R.A., FILIPPI, L.S., CHIAPPE, L.M., GARCÍA, R. & ARCUCCI, A.B. 2013.

1654 Overosaurus paradasorum gen. et sp. nov., a new sauropod dinosaur (Titanosauria:

1655 Lithostrotia) from the Late Cretaceous of Neuquén, Patagonia, Argentina. Zootaxa

1656 3683, 357–376.

1657 CSIKI, Z., CODREA, V., JIPA-MURZEA, C. & GODEFROIT, P. 2010. A partial titanosaur

1658 (Sauropoda, Dinosauria) skeleton from the Maastrichtian of Nǎlaţ-Vad, Haţeg Basin,

1659 Romania. Neues Jahrbuch für Geologie und Paläontologie, Abhandlungen 258, 297–

1660 324.

1661 CURRY ROGERS, K. 2009. The postcranial osteology of Rapetosaurus krausei (Sauropoda:

1662 Titanosauria) from the Late Cretaceous of Madagascar. Journal of Vertebrate

1663 Paleontology 29, 1046–1086.

1664 D'EMIC, M.D. 2012. Early evolution of titanosauriform sauropod dinosaurs. Zoological

1665 Journal of the Linnean Society 166, 624–671.

1666 D'EMIC, M.D. 2013. Revision of the sauropod dinosaurs of the Lower Cretaceous Trinity

1667 Group, southern USA, with the description of a new genus. Journal of Systematic

1668 Palaeontology 11, 707–726.

1669 D'EMIC, M.D. & FOREMAN, B.Z. 2012. The beginning of the sauropod hiatus in North

1670 America: insights from the Cloverly Formation of Wyoming. Journal of Vertebrate

1671 Paleontology 32, 883–902.

1672 D'EMIC, M.D., WILSON, J.A. & CHATTERJEE, S. 2009. The titanosaur (Dinosauria:

1673 Sauropoda) osteoderm record: review and first definitive specimen from India.

1674 Journal of Vertebrate Paleontology 29, 165–177.

1675 DANISE, S. & HIGGS, N.D. 2015. Bone-eating Osedax worms lived on Mesozoic marine

1676 reptile deadfalls. Biology Letters 11, 20150072.

1677 DAVIS, P.G. & BRIGGS, D.E.G. 1998. The impact of decay and disarticulation on the

1678 preservation of fossil birds. PALAIOS 13, 3–13.

1679 DAY, R.W., 1969. The Lower Cretaceous of the Great Artesian Basin. In Stratigraphy and

1680 Palaeontology: Essays in Honour of Dorothy Hill. CAMPBELL, K.S.W., ed,

1681 Australian National University Press, Canberra, 140–173.

1682 DE JESUS FARIA, C.C., GONZÁLEZ RIGA, B., CANDEIRO, C.R.D.A., MARINHO,

1683 T.D.S., ORTIZ DAVID, L., SIMBRAS, F.M., CASTANHO, R.B., MUNIZ, F.P. & DA

1684 COSTA PEREIRA, P.V.L.G. 2015. Cretaceous sauropod diversity and taxonomic

1685 succession in South America. Journal of South American Earth Sciences 61, 154–163.

1686 DÍEZ DÍAZ, V., MOCHO, P., PÁRAMO, M.E., ESCASO, F., MARCOS-FERNÁNDEZ, F.,

1687 SANZ, J.L. & ORTEGA, F. 2016. A new titanosaur (Dinosauria, Sauropoda) from the

1688 Upper Cretaceous of Lo Hueco (Cuenca, Spain). Cretaceous Research 68, 49–60.

1689 DÍEZ DÍAZ, V., PEREDA SUBERBIOLA, X. & SANZ, J.L. 2013. The axial skeleton of the

1690 titanosaur Lirainosaurus astibiae (Dinosauria: Sauropoda) from the latest Cretaceous

1691 of Spain. Cretaceous Research 43, 145–160.

1692 ETHERIDGE, R., JUNR. 1892. Note on Queensland Cretaceous Crustacea. Proceedings of

1693 the Linnean Society of New South Wales 7 (Second Series), 305–306.

1694 ETHERIDGE, R., JUNR. 1917. Descriptions of some Queensland Palaeozoic and Mesozoic

1695 fossils, 1. Queensland Lower Cretaceous Crustacea. Publications of the Geological

1696 Survey of Queensland 260, 1–28.

1697 EXON, N.F. & SENIOR, B.R. 1976. The Cretaceous of the Eromanga and Surat Basins.

1698 Bureau of Mineral Resources, Geology and Geophysics Bulletin 1, 33–50.

1699 FILIPPINI, F.S., OTERO, A. & GASPARINI, Z. 2017. The phylogenetic relevance of the

1700 sacrum among macronarian sauropods: insights from a pelvis from the Upper

1701 Cretaceous of Patagonia, Argentina. Alcheringa 41, 69–78.

1702 FOSTER, J., 2007. Jurassic West: The Dinosaurs of the Morrison Formation and Their

1703 World. Indiana University Press, Bloomington. 387 pp.

1704 FOWLER, D.W. & SULLIVAN, R.M. 2011. The first giant titanosaurian sauropod from the

1705 Upper Cretaceous of North America. Acta Palaeontologica Polonica 56, 685–690.

1706 GAFFNEY, E.S. 1981. A review of the fossil turtles of Australia. American Museum

1707 Novitates 2720, 1–38.

1708 GALLINA, P.A. 2016. Reappraisal of the Early Cretaceous sauropod dinosaur Amargatitanis

1709 macni (Apesteguía, 2007), from northwestern Patagonia, Argentina. Cretaceous

1710 Research 64, 79–87.

1711 GHILARDI, A.M., AURELIANO, T., DUQUE, R.R.C., FERNANDES, M.A., BARRETO,

1712 A.M.F. & CHINSAMY, A. 2016. A new titanosaur from the Lower Cretaceous of

1713 Brazil. Cretaceous Research 67, 16–24.

1714 GLAESSNER, M.F. 1980. New Cretaceous and Tertiary crabs (Crustacea: Brachyura) from

1715 Australia and New Zealand. Transactions of the Royal Society of South Australia 104,

1716 171–192.

1717 GOMANI, E.M. 2005. Sauropod dinosaurs from the Early Cretaceous of Malawi, Africa.

1718 Palaeontologia Electronica 8, 27A.

1719 GONZÁLEZ RIGA, B.J., LAMANNA, M.C., ORTIZ DAVID, L.D., CALVO, J.O. &

1720 CORIA, J.P. 2016. A gigantic new dinosaur from Argentina and the evolution of the

1721 sauropod hind foot. Scientific Reports 6, 19165.

1722 GORSCAK, E. & O'CONNOR, P.M. 2016. Time-calibrated models support congruency

1723 between Cretaceous continental rifting and titanosaurian evolutionary history. Biology

1724 Letters 12, 20151047.

1725 GORSCAK, E., O'CONNOR, P.M., STEVENS, N.J. & ROBERTS, E.M. 2014. The basal

1726 titanosaurian Rukwatitan bisepultus (Dinosauria, Sauropoda) from the middle

1727 Cretaceous Galula Formation, Rukwa Rift Basin, southwestern Tanzania. Journal of

1728 Vertebrate Paleontology 34, 1133–1154.

1729 GRAY, A.R.G., MCKILLOP, M. & MCKELLAR, J.L., 2002. Eromanga Basin Stratigraphy.

1730 In Geology of the Cooper and Eromanga Basins, Queensland. DRAPER, J.J., ed,

1731 Department of Natural Resources and Mines, Brisbane, 30–56.

1732 GREENTREE, J.S., 2011. Palaeoenvironmental setting of dinosaur trackways in the context

1733 of the closing stages of Eromanga Basin evolution. BSc (Hons) thesis. Queensland

1734 University of Technology, Brisbane, Australia. 107 pp.

1735 HAIG, D.W. 1979A. Cretaceous foraminiferal biostratigraphy of Queensland. Alcheringa 3,

1736 171–187.

1737 HAIG, D.W. 1979B. Global distribution patterns for mid-Cretaceous foraminiferids. Journal

1738 of Foraminiferal Research 9, 29–40.

1739 HAIG, D.W. 1980. Early Cretaceous textulariine foraminiferids from Queensland.

1740 Palaeontographica Abteilung A 170, 87–138.

1741 HAIG, D.W. 1982. Early Cretaceous milioline and rotaliine benthic foraminiferids from

1742 Queensland. Palaeontographica Abteilung A 177, 1–88.

1743 HAIG, D.W. & LYNCH, D.A. 1993. A late early Albian marine transgressive pulse over

1744 northeastern Australia, precursor to epeiric basin anoxia: Foraminiferal evidence.

1745 Marine Micropaleontology 22, 311–362.

1746 HATCHER, J.B. 1901. Diplodocus (Marsh): its osteology, taxonomy and probable habits,

1747 with a restoration of the skeleton. Memoirs of the Carnegie Museum 1, 1–63.

1748 HENDERSON, D.M. 2004A. Tipsy punters: sauropod dinosaur pneumaticity, buoyancy and

1749 aquatic habits. Proceedings of the Royal Society B 271, S180–S183.

1750 HENDERSON, R.A. 2004B. A mid-Cretaceous association of shell beds and organic-rich

1751 shale: bivalve exploitation of a nutrient-rich, anoxic sea-floor environment. PALAIOS

1752 19, 156–169.

1753 HENDERSON, R.A. & KENNEDY, W.J. 2002. Occurrence of the ammonite Goodhallites

1754 goodhalli (J. Sowerby) in the Eromanga Basin, Queensland: an index species for the

1755 late Albian. Alcheringa 26, 233–247.

1756 HENDERSON, R.A. & MCKENZIE, E.D. 2002. Idanoceras, a new heteromorph ammonite

1757 genus from the late Albian of Eastern Australia. Journal of Paleontology 76, 906–

1758 909.

1759 HOCKNULL, S.A., WHITE, M.A., TISCHLER, T.R., COOK, A.G., CALLEJA, N.D.,

1760 SLOAN, T. & ELLIOTT, D.A. 2009. New mid-Cretaceous (latest Albian) dinosaurs

1761 from Winton, Queensland, Australia. PLoS ONE 4, e6190.

1762 HOGLER, J.A. 1994. Speculations on the role of marine reptile deadfalls in Mesozoic deep-

1763 sea paleoecology. PALAIOS 9, 42–47.

1764 HOLLAND, T., 2015. New material of Kronosaurus queenslandicus (Plesiosauria,

1765 Pliosauridae) from the Early Cretaceous of Nelia, Queensland. In 15th Conference on

1766 Australasian Vertebrate Evolution, Palaeontology and Systematics. Alice Springs,

1767 Australia, 28–29.

1768 HUENE, F.V. 1929. Los Saurisquios y Ornitisquios del Cretáceo Argentina. Anales Museo de

1769 La Plata 3, 1–196.

1770 HUENE, F.V. 1932. Die fossile Reptil-Ordnung Saurischia, ihre Entwicklung und

1771 Geschichte. Monographien zur Geologie und Paläontologie (Series 1) 4, 1–361.

1772 HUGHES, J. 2003. Marathon Run: The story behind the discovery of the Richmond Pliosaur.

1773 Australian Age of Dinosaurs Museum of Natural History Annual 1, 30–34.

1774 IBIRICU, L.M., CASAL, G.A., MARTÍNEZ, R.D., LAMANNA, M.C., LUNA, M. &

1775 SALGADO, L. 2013. Katepensaurus goicoecheai, gen. et sp. nov., a Late Cretaceous

1776 rebbachisaurid (Sauropoda, Diplodocoidea) from central Patagonia, Argentina.

1777 Journal of Vertebrate Paleontology 33, 1351–1366.

1778 JANENSCH, W. 1947. Pneumatizität bei Wirbeln von Sauropoden und anderen Saurischien.

1779 Palaeontographica, Supplement VII 3, 1–25.

1780 JANENSCH, W. 1950. Die Wirbelsäule von Brachiosaurus brancai. Palaeontographica,

1781 Supplement VII 3, 27–93.

1782 JELL, J.A., COOK, A.G. & JELL, P.A. 2011. Australian Cretaceous Cnidaria and Porifera.

1783 Alcheringa 35, 241–284.

1784 KAIM, A., KOBAYASHI, Y., ECHIZENYA, H., JENKINS, R.G. & TANABE, K. 2008.

1785 Chemosynthesis-based associations on Cretaceous plesiosaurid carcasses. Acta

1786 Palaeontologica Polonica 53, 97–104.

1787 KEAR, B.P. 2001A. Dental caries in an Early Cretaceous ichthyosaur. Alcheringa 25, 387–

1788 390.

1789 KEAR, B.P. 2001B. Elasmosaur (Reptilia: Sauropterygia) basicranial remains from the Early

1790 Cretaceous of Queensland. Records of the South Australian Museum 34, 127–133.

1791 KEAR, B.P. 2003. Cretaceous marine reptiles of Australia: a review of taxonomy and

1792 distribution. Cretaceous Research 24, 277–303.

1793 KEAR, B.P. 2005. Cranial morphology of Platypterygius longmani Wade, 1990 (Reptilia:

1794 Ichthyosauria) from the Lower Cretaceous of Australia. Zoological Journal of the

1795 Linnean Society 145, 583–622.

1796 KEAR, B.P. 2006. First gut contents in a Cretaceous sea turtle. Biology Letters 2, 113–115.

1797 KEAR, B.P. & LEE, M.S.Y. 2006. A primitive protostegid from Australia and early sea turtle

1798 evolution. Biology Letters 2, 116–119.

1799 KEAR, B.P. & ZAMMIT, M. 2014. In utero foetal remains of the Cretaceous ichthyosaurian

1800 Platypterygius: ontogenetic implications for character state efficacy. Geological

1801 Magazine 151, 71–86.

1802 KEMP, N.R., 1991. Chondrichthyans in the Cretaceous and Tertiary of Australia. In

1803 Vertebrate Palaeontology of Australasia. VICKERS-RICH, P., MONAGHAN, J.M.,

1804 BAIRD, R.F. & RICH, T.H., eds, Pioneer Design Studio, Melbourne, 497–568.

1805 KRÖMMELBEIN, K. 1975. Ostracoden aus der Kreide des Great Artesian Basin,

1806 Queensland, Australien. Senckenbergiana lethaea 55, 455–483.

1807 LACOVARA, K.J., LAMANNA, M.C., IBIRICU, L.M., POOLE, J.C., SCHROETER, E.R.,

1808 ULLMANN, P.V., VOEGELE, K.K., BOLES, Z.M., CARTER, A.M., FOWLER,

1809 E.K., EGERTON, V.M., MOYER, A.E., COUGHENOUR, C.L., SCHEIN, J.P.,

1810 HARRIS, J.D., MARTÍNEZ, R.D. & NOVAS, F.E. 2014. A gigantic, exceptionally

1811 complete titanosaurian sauropod dinosaur from southern Patagonia, Argentina.

1812 Scientific Reports 4, 6196.

1813 LEAHEY, L.G., MOLNAR, R.E., CARPENTER, K., WITMER, L.M. & SALISBURY, S.W.

1814 2015. Cranial osteology of the ankylosaurian dinosaur formerly known as Minmi sp.

1815 (Ornithischia: Thyreophora) from the Lower Cretaceous Allaru Mudstone of

1816 Richmond, Queensland, Australia. PeerJ 3, e1475.

1817 LEAHEY, L.G. & SALISBURY, S.W. 2013. First evidence of ankylosaurian dinosaurs

1818 (Ornithischia: Thyreophora) from the mid-Cretaceous (late Albian–Cenomanian)

1819 Winton Formation of Queensland, Australia. Alcheringa 37, 249–257.

1820 LEES, T. 1986. Catalogue of type, figured and mentioned fossil fish, amphibians and reptiles

1821 held by the Queensland Museum. Memoirs of the Queensland Museum 22, 265–288.

1822 LEES, T.A. & BARTHOLOMAI, A. 1987. Study of a Lower Cretaceous actinopterygian

1823 (Class Pisces) Cooyoo australis from Queensland, Australia. Memoirs of the

1824 Queensland Museum 25, 177–192.

1825 LONG, J.A. 1992. First dinosaur bones from Western Australia. The Beagle, Records of the

1826 Northern Territory Museum of Arts and Sciences 9, 21–27.

1827 LONG, J.A., 1998. Dinosaurs of Australia and New Zealand and other Prehistoric Animals

1828 of the Mesozoic Era. University of New South Wales Press, Sydney. 188 pp.

1829 LONGMAN, H., 1949. Even the giant dinosaurs had to put up with floods. The Courier-Mail

1830 August 13, 1949, 2.

1831 LONGMAN, H.A. 1913. Notes on Portheus australis. Memoirs of the Queensland Museum

1832 2, 94–95.

1833 LONGMAN, H.A. 1916. A tentative Dicynodontia from north western Queensland.

1834 Proceedings of the Royal Society of Queensland 27, ix.

1835 LONGMAN, H.A. 1924. A new gigantic marine reptile from the Queensland Cretaceous,

1836 Kronosaurus queenslandicus new genus and species. Memoirs of the Queensland

1837 Museum 8, 26–28.

1838 LONGMAN, H.A. 1926. A giant dinosaur from Durham Downs, Queensland. Memoirs of the

1839 Queensland Museum 8, 183–194.

1840 LONGMAN, H.A. 1927A. Australia’s largest fossil. The Rhoetosaurus dinosaur. Australian

1841 Museum Magazine 3, 97–102.

1842 LONGMAN, H.A. 1927B. The giant dinosaur: Rhoetosaurus brownei. Memoirs of the

1843 Queensland Museum 9, 1–18.

1844 LONGMAN, H.A. 1929. Palaeontological notes. Memoirs of the Queensland Museum 9,

1845 249–250.

1846 LONGMAN, H.A. 1930. Kronosaurus queenslandicus. A gigantic Cretaceous pliosaur.

1847 Memoirs of the Queensland Museum 10, 1–7.

1848 LONGMAN, H.A. 1932. A new Cretaceous fish. Memoirs of the Queensland Museum 10,

1849 89–97.

1850 LONGMAN, H.A. 1933. A new dinosaur from the Queensland Cretaceous. Memoirs of the

1851 Queensland Museum 10, 131–144.

1852 LONGMAN, H.A. 1943. Further notes on Australian ichthyosaurs. Memoirs of the

1853 Queensland Museum 12, 101–104.

1854 MANNION, P.D., ALLAIN, R. & MOINE, O. 2017. The earliest known titanosauriform

1855 sauropod dinosaur and the evolution of Brachiosauridae. PeerJ 5, e3217.

1856 MANNION, P.D. & BARRETT, P.M. 2013. Additions to the sauropod dinosaur fauna of the

1857 Cenomanian (early Late Cretaceous) Kem Kem beds of Morocco:

1858 palaeobiogeographical implications of the mid-Cretaceous African sauropod fossil

1859 record. Cretaceous Research 45, 49–59.

1860 MANNION, P.D. & CALVO, J.O. 2011. Anatomy of the basal titanosaur (Dinosauria,

1861 Sauropoda) Andesaurus delgadoi from the mid-Cretaceous (Albian–early

1862 Cenomanian) Río Limay Formation, Neuquén Province, Argentina: implications for

1863 titanosaur systematics. Zoological Journal of the Linnean Society 163, 155–181.

1864 MANNION, P.D. & UPCHURCH, P. 2010. A quantitative analysis of environmental

1865 associations in sauropod dinosaurs. Paleobiology 36, 253–282.

1866 MANNION, P.D., UPCHURCH, P., BARNES, R.N. & MATEUS, O. 2013. Osteology of the

1867 Late Jurassic Portuguese sauropod dinosaur Lusotitan atalaiensis (Macronaria) and

1868 the evolutionary history of basal titanosauriforms. Zoological Journal of the Linnean

1869 Society 168, 98–206.

1870 MARSH, O.C. 1878. Principal characters of American Jurassic dinosaurs: Part I. American

1871 Journal of Science 16 (series 3), 411–416.

1872 MARTILL, D.M. 1988. A review of the terrestrial vertebrate fauna of the Oxford Clay

1873 (Callovian–Oxfordian) of England. Mercian Geologist 11, 171–190.

1874 MARTIN, V., SUTEETHORN, V. & BUFFETAUT, E. 1999. Description of the type and

1875 referred material of Phuwiangosaurus sirindhornae Martin, Buffetaut and Suteethorn,

1876 1994, a sauropod from the Lower Cretaceous of Thailand. Oryctos 2, 39–91.

1877 MARTINELLI, A.G., GARRIDO, A.C., FORASIEPI, A.M., PAZ, E.R. & GUROVICH, Y.

1878 2007. Notes on fossil remains from the Early Cretaceous Lohan Cura Formation,

1879 Neuquén Province, Argentina. Gondwana Research 11, 537–552.

1880 MARTÍNEZ, R., GIMÉNEZ, O., RODRÍGUEZ, J., LUNA, M. & LAMANNA, M.C. 2004.

1881 An articulated specimen of the basal titanosaurian (Dinosauria: Sauropoda)

1882 Epachthosaurus sciuttoi from the early Late Cretaceous Bajo Barreal Formation of

1883 Chubut Province, Argentina. Journal of Vertebrate Paleontology 24, 107–120.

1884 MARTÍNEZ, R.D.F., LAMANNA, M.C., NOVAS, F.E., RIDGELY, R.C., CASAL, G.A.,

1885 MARTÍNEZ, J.E., VITA, J.R. & WITMER, L.M. 2016. A basal lithostrotian

1886 titanosaur (Dinosauria: Sauropoda) with a complete skull: implications for the

1887 evolution and paleobiology of Titanosauria. PLoS ONE 11, e0151661.

1888 MATHER, P. 1986. A Time for a Museum: The History of the Queensland Museum 1862–

1889 1986. Memoirs of the Queensland Museum 24, 1–366.

1890 MCHENRY, C., COOK, A.G. & WROE, S. 2005. Bottom-feeding plesiosaurs. Science 310,

1891 75.

1892 MCHENRY, C.R., 2009. ‘Devourer of Gods’: The palaeoecology of the Cretaceous pliosaur

1893 Kronosaurus queenslandicus. Ph.D. thesis. University of Newcastle.

1894 MCINTOSH, J.S., 1990. Sauropoda. In The Dinosauria. WEISHAMPEL, D.B., DODSON, P.

1895 & OSMÓLSKA, H., eds, University of California Press, Berkeley, 345–401.

1896 MCNAMARA, K.J. 1978. Myloceras (Ammonoidea) from the Albian of central Queensland.

1897 Alcheringa 2, 231–242.

1898 MOBBS, C. 1990. The creature from our inland sea. Australian Geographic 20, 26–27.

1899 MOLNAR, R. 1980. Australian late Mesozoic terrestrial tetrapods: some implications.

1900 Mémoires de la Société Géologique de France (Nouvelle Série) 139, 131–143.

1901 MOLNAR, R., 1984A. Palaeozoic and Mesozoic reptiles and amphibians from Australia. In

1902 Vertebrate Zoogeography & Evolution in Australasia (Animals in Space and Time).

1903 ARCHER, M. & CLAYTON, G., eds, Hesperian Press, Sydney, Australia, 331–336.

1904 MOLNAR, R.E., 1982A. Australian Mesozoic reptiles. In Fossil Vertebrate Record of

1905 Australasia. RICH, P.V. & THOMPSON, E.M., eds, Monash University Press,

1906 Clayton, Victoria, 169–225.

1907 MOLNAR, R.E. 1982B. A catalogue of fossil amphibians and reptiles in Queensland.

1908 Memoirs of the Queensland Museum 20, 613–633.

1909 MOLNAR, R.E., 1984B. Ornithischian dinosaurs in Australia. In Third Symposium on

1910 Mesozoic Terrestrial Ecosystems, Short Papers. REIF, W.-E. & WESTPHAL, F., eds,

1911 University Press, Tübingen, 151–156.

1912 MOLNAR, R.E., 1991. Fossil reptiles in Australia. In Vertebrate Palaeontology of

1913 Australasia. VICKERS-RICH, P., MONAGHAN, J.M., BAIRD, R.F. & RICH, T.H.,

1914 eds, Pioneer Design Studio, Melbourne, 605–702.

1915 MOLNAR, R.E. 1996A. Observations on the Australian ornithopod Muttaburrasaurus.

1916 Memoirs of the Queensland Museum 39, 639–652.

1917 MOLNAR, R.E. 1996B. Preliminary report on a new ankylosaur from the Early Cretaceous of

1918 Queensland, Australia. Memoirs of the Queensland Museum 39, 653–668.

1919 MOLNAR, R.E., 2001A. Armor of the small ankylosaur Minmi. In The Armored Dinosaurs.

1920 CARPENTER, K., ed, Indiana University Press, Bloomington and Indianapolis, 341–

1921 362.

1922 MOLNAR, R.E., 2001B. A reassessment of the phylogenetic position of Cretaceous sauropod

1923 dinosaurs from Queensland, Australia. In VII International Symposium on Mesozoic

1924 Terrestrial Ecosystems: Asociacíon Paleontológica Argentina Publicación Especial

1925 No. 7. LEANZA, H.A., ed, Asociacion Paleontologica Argentina, Buenos Aires, 139–

1926 144.

1927 MOLNAR, R.E. 2010. Taphonomic observations on eastern Australian Cretaceous

1928 sauropods. Alcheringa 34, 421–429.

1929 MOLNAR, R.E. 2011A. New morphological information about Cretaceous sauropod

1930 dinosaurs from the Eromanga Basin, Queensland, Australia. Alcheringa 35, 329–339.

1931 MOLNAR, R.E. 2011B. Sauropod (Saurischia: Dinosauria) material from the Early

1932 Cretaceous Griman Creek Formation of the Surat Basin, Queensland, Australia.

1933 Alcheringa 35, 303–307.

1934 MOLNAR, R.E. & CLIFFORD, H.T. 2000. Gut contents of a small ankylosaur. Journal of

1935 Vertebrate Paleontology 20, 194–196.

1936 MOLNAR, R.E. & CLIFFORD, H.T., 2001. An ankylosaurian cololite from the Lower

1937 Cretaceous of Queensland, Australia. In The Armored Dinosaurs. CARPENTER, K.,

1938 ed, Indiana University Press, Bloomington and Indianapolis, 399–412.

1939 MOLNAR, R.E. & SALISBURY, S.W., 2005. Observations on Cretaceous sauropods from

1940 Australia. In Thunder-lizards: The Sauropodomorph Dinosaurs. TIDWELL, V. &

1941 CARPENTER, K., eds, Indiana University Press, Bloomington & Indianapolis, 454–

1942 465.

1943 NAIR, J.P. & SALISBURY, S.W. 2012. New anatomical information on Rhoetosaurus

1944 brownei Longman, 1926, a gravisaurian sauropodomorph dinosaur from the Middle

1945 Jurassic of Queensland, Australia. Journal of Vertebrate Paleontology 32, 369–394.

1946 NOVAS, F.E., SALGADO, L., CALVO, J.O. & AGNOLÍN, F.L. 2005. Giant titanosaur

1947 (Dinosauria, Sauropoda) from the Late Cretaceous of Patagonia. Revista del Museo

1948 Argentino de Ciencias Naturales 7, 37–41.

1949 OTERO, A., CANALE, J.I., HALUZA, A. & CALVO, J.O. 2011. New titanosaur with

1950 unusual haemal arches from the Upper Cretaceous of Neuquén Province, Argentina.

1951 Ameghiniana 48, 655–661.

1952 OWEN, R. 1842. Report on British fossil reptiles. Part II. Report of the Eleventh Meeting of

1953 the British Association for the Advancement of Science, held at Plymouth, July 1841,

1954 60–204.

1955 OWEN, R. 1876. Monograph on the Fossil Reptilia of the Wealden and Purbeck Formations.

1956 Supplement No. VII. Crocodilia (Poikilopleuron) and Dinosauria?

1957 (Chondrosteosaurus) [Wealden.]. Monograph of the Palaeontographical Society 30,

1958 1–7.

1959 PERSSON, P.O. 1960. Lower Cretaceous plesiosaurians (Rept.) from Australia. Lunds

1960 Universitets Årsskrift 56, 1–23.

1961 PLAYFORD, G., HAIG, D.W. & DETTMANN, M.E. 1975. A mid-Cretaceous microfossil

1962 assemblage from the Great Artesian Basin, northwestern Queensland. Neues Jahrbuch

1963 für Geologie und Paläontologie, Abhandlungen 149, 333–362.

1964 POROPAT, S. 2016. Eight decades of Austrosaurus: The discovery and rediscovery of

1965 Queensland’s first Cretaceous dinosaur. Australian Age of Dinosaurs Museum of

1966 Natural History Annual 13, 24–39.

1967 POROPAT, S.F., MANNION, P.D., UPCHURCH, P., HOCKNULL, S.A., KEAR, B.P. &

1968 ELLIOTT, D.A. 2015A. Reassessment of the non-titanosaurian somphospondylan

1969 Wintonotitan wattsi (Dinosauria: Sauropoda: Titanosauriformes) from the mid-

1970 Cretaceous Winton Formation, Queensland, Australia. Papers in Palaeontology 1,

1971 59–106.

1972 POROPAT, S.F., MANNION, P.D., UPCHURCH, P., HOCKNULL, S.A., KEAR, B.P.,

1973 KUNDRÁT, M., TISCHLER, T.R., SLOAN, T., SINAPIUS, G.H.K., ELLIOTT, J.A.

1974 & ELLIOTT, D.A. 2016. New Australian sauropods shed light on Cretaceous

1975 dinosaur palaeobiogeography. Scientific Reports 6, 34467.

1976 POROPAT, S.F., THULBORN, T. & KEAR, B.P., 2014. A dinosaur lost at sea: a new

1977 titanosauriform sauropod from Lower Cretaceous marine deposits in Australia. In 4th

1978 International Palaeontological Congress Abstract Volume. Mendoza, Argentina, 331.

1979 POROPAT, S.F., UPCHURCH, P., MANNION, P.D., HOCKNULL, S.A., KEAR, B.P.,

1980 SLOAN, T., SINAPIUS, G.H.K. & ELLIOTT, D.A. 2015B. Revision of the sauropod

1981 dinosaur Diamantinasaurus matildae Hocknull et al. 2009 from the middle

1982 Cretaceous of Australia: implications for Gondwanan titanosauriform dispersal.

1983 Gondwana Research 27, 995–1033.

1984 POWELL, J.E. 1990. Epachthosaurus sciuttoi gen. et sp. nov, un nuevo dinosaurio

1985 titanosáurido del Cretácico de Patagonia (Provincia de Chubut, Argentina). Actas del

1986 V Congreso Argentino de Paleontología y Bioestratigrafía, Tucumán, 1990 1, 123–

1987 128.

1988 POWELL, J.E., 1992. Osteologia de Saltasaurus loricatus (Sauropoda–Titanosauridae) del

1989 Cretácico Superior del noroeste Argentino. In Los Dinosaurios y Su Entorno Biótico:

1990 Actas del Segundo Curso de Paleontologia in Cuenca. SANZ, J.L. & BUSCALIONI,

1991 A.D., eds, Instituto "Juan de Valdés", Cuenca, Spain, 165–230.

1992 POWELL, J.E. 2003. Revision of South American titanosaurid dinosaurs: palaeobiological,

1993 palaeobiogeographical and phylogenetic aspects. Records of the Queen Victoria

1994 Museum 111, 1–173.

1995 PRICE, G.D., WILLIAMSON, T., HENDERSON, R.A. & GAGAN, M.K. 2012. Barremian–

1996 Cenomanian palaeotemperatures for Australian seas based on new oxygen-isotope

1997 data from belemnite rostra. Palaeogeography, Palaeoclimatology, Palaeoecology

1998 358–360, 27–39.

1999 RAUHUT, O.W.M., CLADERA, G., VICKERS-RICH, P. & RICH, T.H. 2003. Dinosaur

2000 remains from the Lower Cretaceous of the Chubut Group, Argentina. Cretaceous

2001 Research 24, 487–497.

2002 REISDORF, A.G. & WUTTKE, M. 2012. Re-evaluating Moodie’s opisthotonic-posture

2003 hypothesis in fossil vertebrates Part I: reptiles — the taphonomy of the bipedal

2004 dinosaurs Compsognathus longipes and Juravenator starki from the Solnhofen

2005 Archipelago (Jurassic, Germany). Palaeobiodiversity and Palaeoenvironments 92,

2006 119–168.

2007 RICH, T.H. & VICKERS-RICH, P., 2003. A Century of Australian Dinosaurs. Octavo,

2008 Launceston, Tasmania. 124 pp.

2009 ROGERS, R.R. & KIDWELL, S.M., 2007. A conceptual framework for the genesis and

2010 analysis of vertebrate skeletal concentrations. In Bonebeds: Genesis, Analysis, and

2011 Paleobiological Significance. ROGERS, R.R., EBERTH, D.A. & FIORILLO, A.R.,

2012 eds, University of Chicago Press, Chicago, Illinois, 1–63.

2013 ROMER, A.S. & LEWIS, A.D. 1959. A mounted skeleton of the giant plesiosaur

2014 Kronosaurus. Breviora 112, 1–15.

2015 ROZEFELDS, A. 1986. Type, figured and mentioned fossil plants in the Queensland

2016 Museum. Memoirs of the Queensland Museum 22, 141–153.

2017 SALGADO, L. & BONAPARTE, J.F., 2007. Sauropodomorpha. In Patagonian Mesozoic

2018 Reptiles. GASPARINI, Z., SALGADO, L. & CORIA, R.A., eds, Indiana University

2019 Press, Bloomington & Indianapolis, 188–228.

2020 SALGADO, L. & CORIA, R.A., 2005. Sauropods of Patagonia: systematic update and notes

2021 on global sauropod evolution. In Thunder-lizards: The Sauropodomorph Dinosaurs.

2022 TIDWELL, V. & CARPENTER, K., eds, Indiana University Press, Bloomington &

2023 Indianapolis, 430–453.

2024 SALGADO, L. & CORIA, R.A. 2009. Barrosasaurus casamiquelai gen. et sp. nov., a new

2025 titanosaur (Dinosauria, Sauropoda) from the Anacleto Formation (Late Cretaceous:

2026 early Campanian) of Sierra Barrosa (Neuquén, Argentina). Zootaxa 2222, 1–16.

2027 SALGADO, L., CORIA, R.A. & CALVO, J.O. 1997. Evolution of titanosaurid sauropods. I:

2028 Phylogenetic analysis based on the postcranial evidence. Ameghiniana 34, 3–32.

2029 SALISBURY, S.W., ROMILIO, A., HERNE, M.C., TUCKER, R.T. & NAIR, J.P. 2017. The

2030 dinosaurian ichnofauna of the Lower Cretaceous (Valanginian–Barremian) Broome

2031 Sandstone of the Walmadany Area (James Price Point), Dampier Peninsula, Western

2032 Australia. Society of Vertebrate Paleontology Memoir 16 (Journal of Vertebrate

2033 Paleontology Vol. 36, supplement to No. 6, November 2016), i–viii + 152 pp.

2034 SANTUCCI, R.M. & ARRUDA-CAMPOS, A.C. 2011. A new sauropod (Macronaria,

2035 Titanosauria) from the Adamantina Formation, Bauru Group, Upper Cretaceous of

2036 Brazil and the phylogenetic relationships of Aeolosaurini. Zootaxa 3085, 1–33.

2037 SCHÄFER, W., ed (1972) Ecology and Palaecology of Marine Environments. The University

2038 of Chicago Press, Chicago.

2039 SEELEY, H.G. 1870. On Ornithopsis, a gigantic animal of the pterodactyle kind from the

2040 Wealden. Annals and Magazine of Natural History (Series 4) 5, 279–283.

2041 SEELEY, H.G. 1887. On the classification of the fossil animals commonly named

2042 Dinosauria. Proceedings of the Royal Society of London 43, 165–171.

2043 SETON, M., MÜLLER, R.D., ZAHIROVIC, S., GAINA, C., TORSVIK, T., SHEPARD, G.,

2044 TALSMA, A., GURNIS, M., TURNER, M., MAUS, S. & CHANDLER, M. 2012.

2045 Global continental and ocean basin reconstructions since 200 Ma. Earth-Science

2046 Reviews 113, 212–270.

2047 SHAFIK, S. 1985. Calcareous nannofossils from the Toolebuc Formation, Eromanga Basin,

2048 Australia. Bureau of Mineral Resources Journal of Australian Geology and

2049 Geophysics 9, 171–181.

2050 SMITH, C.R. & BACO, A.R. 2003. Ecology of whale falls at the deep-sea floor.

2051 Oceanography and Marine Biology: An Annual Review 41, 311–354.

2052 SMITH, P.M. & HOLLAND, T., 2016. Cretaceous time capsules: remarkable preservation of

2053 fish and crustaceans inside the bivalve Inoceramus sutherlandi McCoy, 1865 from the

2054 Allaru Mudstone (late Albian), Eromanga Basin, Queensland. In Palaeo Down Under

2055 2, 11–15 July 2016. LAURIE, J.R., KRUSE, P.D., GARCÍA-BELLIDO, D.C. &

2056 HOLMES, J.D., eds, Geological Society of Australia, Adelaide, 55–56.

2057 STILWELL, J.D. 1999. Cretaceous Scaphopoda (Mollusca) of Australia and their

2058 palaeobiogeographic significance. Alcheringa 23, 215–226.

2059 SUTEETHORN, S., LE LOEUFF, J., BUFFETAUT, E. & SUTEETHORN, V. 2010.

2060 Description of topotypes of Phuwiangosaurus sirindhornae, a sauropod from the Sao

2061 Khua Formation (Early Cretaceous) of Thailand, and their phylogenetic implications.

2062 Neues Jahrbuch für Geologie und Paläontologie, Abhandlungen 256, 109–121.

2063 SUTEETHORN, S., LE LOEUFF, J., BUFFETAUT, E., SUTEETHORN, V., TALUMBOOK,

2064 C. & CHONGLAKMANI, C., 2009. A new skeleton of Phuwiangosaurus

2065 sirindhornae (Dinosauria, Sauropoda) from NE Thailand. In Late Palaeozoic and

2066 Mesozoic Ecosystems in SE Asia. The Geological Society, London, Special

2067 Publications, 315. BUFFETAUT, E., CUNY, G., LE LOEUFF, J. & SUTEETHORN,

2068 V., eds, Geological Society, London, 189–215.

2069 SYME, C.E. & SALISBURY, S.W. 2014. Patterns of aquatic decay and disarticulation in

2070 juvenile Indo-Pacific crocodiles (Crocodylus porosus), and implications for the

2071 taphonomic interpretation of fossil crocodyliform material. Palaeogeography,

2072 Palaeoclimatology, Palaeoecology 412, 108–123.

2073 TAYLOR, M.P. 2009. A re-evaluation of Brachiosaurus altithorax Riggs 1903 (Dinosauria,

2074 Sauropoda) and its generic separation from Giraffatitan brancai (Janensch 1914).

2075 Journal of Vertebrate Paleontology 29, 787–806.

2076 THULBORN, R.A., 1985. Rhoetosaurus brownei Longman, 1926: the giant Queensland

2077 dinosaur. In Kadimakara: Extinct Vertebrates of Australia. RICH, P.V., VAN TETS,

2078 G.F. & KNIGHT, F., eds, Princeton University Press, Princeton, New Jersey, 166–

2079 171.

2080 THULBORN, R.A. & WADE, M. 1979. Dinosaur stampede in the Cretaceous of Queensland.

2081 Lethaia 12, 275–279.

2082 THULBORN, R.A. & WADE, M. 1984. Dinosaur trackways in the Winton Formation (mid-

2083 Cretaceous) of Queensland. Memoirs of the Queensland Museum 21, 413–517.

2084 THULBORN, T., 1987. An enigmatic brontosaur: Austrosaurus mckillopi. In The Antipodean

2085 Ark. HAND, S. & ARCHER, M., eds, Angus & Robertson, Sydney, 44–46.

2086 THULBORN, T. 2012. Impact of sauropod dinosaurs on lagoonal substrates in the Broome

2087 Sandstone (Lower Cretaceous), Western Australia. PLoS ONE 7, e36208.

2088 THULBORN, T., HAMLEY, T. & FOULKES, P., 1994. Preliminary report on sauropod

2089 dinosaur tracks in the Broome Sandstone (Lower Cretaceous) of Western Australia. In

2090 Aspects of Sauropod Paleobiology; Gaia, 10. LOCKLEY, M.G., SANTOS, V.F.,

2091 MEYER, C.A. & HUNT, A.P., eds, Lisbon, Portugal, 85–94.

2092 THULBORN, T. & TURNER, S. 2003. The last dicynodont: an Australian relict. Proceedings

2093 of the Royal Society B 270, 985–993.

2094 TIDWELL, V., CARPENTER, K. & BROOKS, W. 1999. New sauropod from the Lower

2095 Cretaceous of Utah, USA. Oryctos 2, 21–37.

2096 TRUEMAN, C.N. & MARTILL, D.M. 2002. The long-term survival of bone: the role of

2097 bioerosion. Archaeometry 44, 371–382.

2098 TUCKER, R.T., 2014. Stratigraphy, sedimentation and age of the Upper Cretaceous Winton

2099 Formation, central-western Queensland, Australia: implications for regional

2100 palaeogeography, palaeoenvironments and Gondwanan palaeontology. Ph.D. thesis.

2101 James Cook University, Townsville, Australia. 209 pp.

2102 TUCKER, R.T., ROBERTS, E.M., HU, Y., KEMP, A.I.S. & SALISBURY, S.W. 2013.

2103 Detrital zircon age constraints for the Winton Formation, Queensland: contextualizing

2104 Australia’s Late Cretaceous dinosaur faunas. Gondwana Research 24, 767–779.

2105 TYKOSKI, R.S. & FIORILLO, A.R. 2017. An articulated cervical series of Alamosaurus

2106 sanjuanensis Gilmore, 1922 (Dinosauria, Sauropoda) from Texas: new perspective on

2107 the relationships of North America's last giant sauropod. Journal of Systematic

2108 Palaeontology 15, 339–364.

2109 UPCHURCH, P. 1995. The evolutionary history of sauropod dinosaurs. Philosophical

2110 Transactions: Biological Sciences 349, 365–390.

2111 UPCHURCH, P. 1998. The phylogenetic relationships of sauropod dinosaurs. Zoological

2112 Journal of the Linnean Society 124, 43–103.

2113 UPCHURCH, P., BARRETT, P.M. & DODSON, P., 2004. Sauropoda. In The Dinosauria:

2114 Second Edition. WEISHAMPEL, D.B., DODSON, P. & OSMÓLSKA, H., eds,

2115 University of California Press, Berkeley, 259–322.

2116 UPCHURCH, P., MANNION, P.D. & TAYLOR, M.P. 2015. The anatomy and phylogenetic

2117 relationships of “Pelorosaurus” becklesii (Neosauropoda, Macronaria) from the Early

2118 Cretaceous of England. PLoS ONE 10, e0125819.

2119 UPCHURCH, P. & MARTIN, J. 2002. The Rutland Cetiosaurus: the anatomy and

2120 relationships of the Middle Jurassic British sauropod dinosaur. Palaeontology 45,

2121 1049–1074.

2122 VINE, R.R., DAY, R.W., MILLIGAN, E.N., CASEY, D.J., GALLOWAY, M.C. & EXON,

2123 N.F. 1967. Revision of the nomenclature of the Rolling Downs Group in the

2124 Eromanga and Surat Basins. Queensland Government Mining Journal 68, 144–151.

2125 WADE, M. 1984. Platypterygius australis, an Australian Cretaceous ichthyosaur. Lethaia 17,

2126 99–113.

2127 WADE, M. 1990. A review of the Australian Cretaceous longipinnate ichthyosaur

2128 Platypterygius, (Ichthyosauria, Ichthyopterygia). Memoirs of the Queensland Museum

2129 28, 115–137.

2130 WADE, M., 1993. New Kelaenida and Vampyromorpha: Cretaceous squid from Queensland.

2131 In Palaeontological Studies in Honour of Ken Campbell. Memoirs of the Association

2132 of Australian Palaeontologists, 15. JELL, P.A., ed, Association of Australasian

2133 Palaeontologists, Brisbane, 353–374.

2134 WASKOW, K. & SANDER, P.M. 2014. Growth record and histological variation in the

2135 dorsal ribs of Camarasaurus sp. (Sauropoda). Journal of Vertebrate Paleontology 34,

2136 852–869.

2137 WEDEL, M.J. 2003. The evolution of vertebral pneumaticity in sauropod dinosaurs. Journal

2138 of Vertebrate Paleontology 23, 344–357.

2139 WEDEL, M.J. 2009. Evidence for bird-like air sacs in saurischian dinosaurs. Journal of

2140 Experimental Zoology Part A: Ecological Genetics and Physiology 311A, 611–628.

2141 WEDEL, M.J., CIFELLI, R.L. & SANDERS, R.K. 2000. Osteology, paleobiology, and

2142 relationships of the sauropod dinosaur Sauroposeidon. Acta Palaeontologica Polonica

2143 45, 343–388.

2144 WHITE, T.E. 1935. On the skull of Kronosaurus queenslandicus Longman. Occasional

2145 Papers of the Boston Society of Natural History 8, 219–228.

2146 WILLIS, P. & THOMAS, A., 2005. Digging up Deep Time: Fossils, Dinosaurs and

2147 Megabeasts from Australia's Distant Past. ABC Books, Sydney, Australia. 294 pp.

2148 WILSON, G.D.F., PATERSON, J.R. & KEAR, B.P. 2011A. Fossil isopods associated with a

2149 fish skeleton from the Lower Cretaceous of Queensland, Australia – direct evidence

2150 of a scavenging lifestyle in Mesozoic Cymothoida. Palaeontology 54, 1053–1068.

2151 WILSON, J.A. 2002. Sauropod dinosaur phylogeny: critique and cladistic analysis.

2152 Zoological Journal of the Linnean Society 136, 217–276.

2153 WILSON, J.A., D'EMIC, M.D., IKEJIRI, T., MOACDIEH, E.M. & WHITLOCK, J.A. 2011B.

2154 A nomenclature for vertebral fossae in sauropods and other saurischian dinosaurs.

2155 PLoS ONE 6, e17114.

2156 WILSON, J.A., POL, D., CARVALHO, A.B. & ZAHER, H. 2016. The skull of the titanosaur

2157 Tapuiasaurus macedoi (Dinosauria: Sauropoda), a basal titanosaur from the Lower

2158 Cretaceous of Brazil. Zoological Journal of the Linnean Society 178, 611–662.

2159 WILSON, J.A. & SERENO, P.C. 1998. Early evolution and higher-level phylogeny of

2160 sauropod dinosaurs. Society of Vertebrate Paleontology Memoir 5 (Journal of

2161 Vertebrate Paleontology Vol. 18, supplement to No. 2, June 1998), 68 pp.

2162 WILSON, J.A. & UPCHURCH, P. 2003. A revision of Titanosaurus Lydekker (Dinosauria –

2163 Sauropoda), the first dinosaur genus with a ‘Gondwanan’ distribution. Journal of

2164 Systematic Palaeontology 1, 125–160.

2165 WILSON, J.A. & UPCHURCH, P. 2009. Redescription and reassessment of Euhelopus

2166 zdanskyi (Dinosauria: Sauropoda) from the Early Cretaceous of China. Journal of

2167 Systematic Palaeontology 7, 199–239.

2168 WOODS, J.T. 1953. Brachyura from the Cretaceous of Queensland. Memoirs of the

2169 Queensland Museum 13, 50–57.

2170 WOODWARD, H. 1892. Note on a new decapodous crustacean, Prosopon etheridgei H.

2171 Woodw. from the Cretaceous of Queensland. Proceedings of the Linnean Society of

2172 New South Wales 7 (Second Series), 301–304.

2173 WRETMAN, L. & KEAR, B.P. 2014. Bite marks on an ichthyodectiform fish from Australia:

2174 possible evidence of trophic interaction in an Early Cretaceous marine ecosystem.

2175 Alcheringa 38, 170–176.

2176 YOU, H.-L., LI, D.-Q., ZHOU, L.-Q. & JI, Q. 2008. Daxiatitan binglingi: a giant sauropod

2177 dinosaur from the Early Cretaceous of China. Gansu Geology 17, 1–10.

2178 ZAHER, H., POL, D., CARVALHO, A.B., NASCIMENTO, P.M., RICCOMINI, C.,

2179 LARSON, P., JUAREZ-VALIERI, R., PIRES-DOMINGUES, R., SILVA, N.J.D., JR.

2180 & CAMPOS, D.D.A. 2011. A complete skull of an Early Cretaceous sauropod and the

2181 evolution of advanced titanosaurians. PLoS ONE 6, e16663.

2182 ZAMMIT, M., DANIELS, C.B. & KEAR, B.P. 2008. Elasmosaur (Reptilia: Sauropterygia)

2183 neck flexibility: Implications for feeding strategies. Comparative Biochemistry and

2184 Physiology - Part A: Molecular & Integrative Physiology 150, 124–130.

2185 ZAMMIT, M., NORRIS, R.M. & KEAR, B.P. 2010. The Australian Cretaceous ichthyosaur

2186 Platypterygius australis: a description and review of postcranial remains. Journal of

2187 Vertebrate Paleontology 30, 1726–1735.

2188

2189 Figure captions

2190 Figure 1: Geological map of the Julia Creek–Richmond region of North West Queensland,

2191 showing the position of Clutha Station (geological outcrop data from Vine et al. 1963, 1970).

2192 Scale bar = 10 km.

2193 Figure 2: Austrosaurus mckillopi, historical overview. A) Henry Burgoyne Wade (1902–

2194 1970; courtesy Peter Wade), B) Harley John McKillop (1888–1967; courtesy Elizabeth

2195 Cleary [née McKillop]), and C) Dr Martin Joseph McKillop (1893–1980; courtesy Elizabeth

2196 Cleary [née McKillop] and Kathryn Evans [née McKillop]) discovered and excavated the

2197 Austrosaurus mckillopi type specimen in the early 1930s. D) Heber Albert Longman (1880–

2198 1954) named Austrosaurus mckillopi on March 14th, 1933 (courtesy Queensland Museum).

2199 E) The gidgee (Acacia) post-supported sign at the Austrosaurus type site in 1933 (courtesy

2200 Peter Wade). Figure 3: A) Map of Clutha Station showing the

2201 Austrosaurus mckillopi type site (marked with an X), supplied by H. J. McKillop to H. A.

2202 Longman, who published it in the original description of Austrosaurus (Longman 1933). B)

2203 Geological map of Clutha Station showing the lithology (Kla = Allaru Mudstone), paddock

2204 fence lines, creeks, and location of the homestead (image compiled from Vine et al. 1963,

2205 1970).

2206 Figure 4: Rearticulated vertebral column of the Austrosaurus mckillopi type specimen (QM

2207 F2316) in left lateral view, with anterior and posterior views of each block. A1) anterior view

2208 of cross-section through the posteriormost cervical vertebra; A2) posterior view of cross-

2209 section through dorsal vertebra I; B1) anterior view of cross-section through dorsal vertebra I;

2210 B2) posterior view of cross-section through dorsal vertebra II; C1) anterior view of cross-

2211 section through dorsal vertebra II; C2) posterior view of cross-section through dorsal vertebra

2212 III; D1) anterior view of cross-section through dorsal vertebra III; D2) posterior view of

2213 cross-section through dorsal vertebra IV; E1) anterior view of cross-section through dorsal

2214 vertebra IV; E2) posterior view of cross-section through dorsal vertebra IV; F1) anterior view

2215 of cross-section through dorsal vertebra V; F2) posterior view of cross-section through dorsal

2216 vertebra V. UCV = ultimate cervical vertebra; DV # = dorsal vertebra #. Scale bar = 200 mm.

2217

2218 Figure 5: Rearticulated vertebral column of the Austrosaurus mckillopi type specimen (QM

2219 F2316) in left lateral view: A) photograph; B) schematic. UCV = ultimate cervical vertebra;

2220 DV # = dorsal vertebra #. Hatched sections on schematic indicate matrix. Scale bar = 100

2221 mm.

2222 Figure 6: Rearticulated vertebral column of the Austrosaurus mckillopi type specimen (QM

2223 F2316) in right lateral view: A) photograph; B) schematic. C) Posterior portion of ultimate

2224 cervical vertebra and anterior portion of dorsal vertebra I showing the autapomorphic lateral

2225 accessory foramen anterior to the lateral pneumatic foramen. UCV = ultimate cervical

2226 vertebra; DV # = dorsal vertebra #. Hatched sections on schematic indicate matrix. Scale bar

2227 = 100 mm.

2228 Figure 7: Rearticulated vertebral column of the Austrosaurus mckillopi type specimen (QM

2229 F2316) in dorsal view: A) photograph with blocks of neural arches in place; B) photograph

2230 with blocks of neural arches removed; C) schematic. UCV = ultimate cervical vertebra; DV #

2231 = dorsal vertebra #. Hatched sections on schematic indicate matrix. Scale bar = 100 mm.

2232

2233 Figure 8: Rearticulated vertebral column of the Austrosaurus mckillopi type specimen (QM

2234 F2316) in ventral view: A) photograph; B) schematic. UCV = ultimate cervical vertebra; DV

2235 # = dorsal vertebra #. Hatched sections on schematic indicate matrix. Scale bar = 100 mm.

2236

2237 Figure 9: Dorsal ribs of the Austrosaurus mckillopi type specimen (KK F1020) in lateral

2238 view, each of which includes several cross-sections. The medial surface of each cross-section

2239 is directed towards the top of the page. DR # = dorsal rib #. Scale bar = 200 mm.

2240 COLUMN WIDTH>

2241 Figure 10: Retrospective map of the Austrosaurus mckillopi type site. The positions of the

2242 ribs (KK F1020) were plotted using hand-drawn site maps and photographs from the 2014

2243 and 2015 excavations, whereas the rearticulated presacral vertebrae (QM F2316) were

2244 emplaced on the basis of their interpreted serial position and information derived from the rib

2245 fragments embedded in matrix adhered to the vertebrae. Scale bar = 200 mm.

2246 WIDTH>

2247 Figure 11: Reconstruction of the possible sequence of events that led to the preservation of

2248 the carcass of the sauropod Austrosaurus mckillopi in the Eromanga Sea. A) Austrosaurus as

2249 a living animal on land; B) freshly deceased Austrosaurus prior to bloating; C) bloated

2250 Austrosaurus carcass washed out to sea, where it was possibly scavenged by marine reptiles

2251 like Kronosaurus; D) the partially defleshed but still effectively intact thoracic portion of the

2252 Austrosaurus carcass is picked at by sharks as it sinks to the seafloor; E) the thoracic portion

2253 of the Austrosaurus carcass is buried along with several ammonites (Beudanticeras) and

2254 bivalves (Inoceramus) which were possibly drawn to the carcass as it decayed.

2255 Reconstruction by Travis R. Tischler.

2256 Figure 12: A, B) Two views of an ammonite (Beudanticeras sp.) preserved within the matrix

2257 adhering to “specimen g”; and C) inoceramid bivalve (Inoceramus sp.) shells in cross-

2258 section, preserved within the matrix adhering to “specimen a” (i.e. between dorsal vertebrae I

2259 and II). Scale bar = 10 mm.

2260 Figure 13: CT rendering of the rearticulated vertebral column of the Austrosaurus mckillopi

2261 type specimen (QM F2316) in A) axial (viewed dorsally) and B) sagittal (viewed left

2262 laterally) mid-line sections. Both the axial and sagittal vertebral series were assembled from

2263 individual CT scan images of each bone (taken through the pneumatic foramina in the case of

2264 the axial section, and through the neural canal in the case of the sagittal section), hence the

2265 imperfect articulation. Scale bar = 100 mm.

2266 Figure 14: Eight of the dorsal ribs from the type specimen (AODF 603) of

2267 Diamantinasaurus matildae. Ribs A, D–F and H are portrayed in lateral view (with the

2268 medial surfaces of the cross-sections directed towards the top of the page); Ribs B and G are

2269 portrayed in medial view (with the medial surfaces of the cross-sections directed towards the

100

2270 bottom of the page); and Rib C is portrayed in posterior view (with the medial surface of the

2271 cross-sections directed towards the left of the page). Scale bar = 200 mm.

2272 WIDTH>

2273 Figure 15: Cervical vertebra of the “Hughenden sauropod” (QM F6142) in A) dorsal; B)

2274 anterior; C) anteroventral; D) left lateral (with schematic below); E) posterior (with schematic

2275 below); F) right lateral (with schematic below); and G) ventral views. H) close-up of the

2276 neural spine in posterior view showing SPOL-Fs. Scale bar at bottom left for A–G = 100 mm;

2277 scale bar at top right for H = 50 mm.

101

2278 Tables

Measurements (mm) Cervical Dorsal vertebrae *Incomplete measurement. vertebra Ultimate I II III IV V Centrum Length (including condyle; ventral 100* 320 260* 280 210* 210* midline) Length (including condyle; left - - - 265 - - lateral) Length (excluding condyle; ventral - 250 230* 217 - - midline) Length (excluding condyle; left - - 185 220 - - lateral) Anterior Maximum width - - - - - 240 (condyle) Posterior Maximum dorsoventral 235 211 - - F: - (cotyle) height (left lateral) 250 Maximum transverse 322 301 - - - - width (ventral) Ratio H:W 0.73 0.7 - - - - Neural canal Maximum transverse breadth - - 67 52 - 52 (anterior) Maximum height of neural canal - - 50 50 - 50 Neural canal Maximum transverse breadth - 54 41 45 42 46 (posterior) Maximum height of neural canal - 43 54 33 47 40 Pneumatic Maximum anteroposterior length - 57 65 110 120 70 foramen Maximum dorsoventral height - 21 45 90 65 70 (left lateral) Pneumatic Maximum anteroposterior length - 90 86 80 115 - foramen Maximum dorsoventral height - 40 53 55 77 - (right lateral)

2279 Table 1. Measurements of the vertebrae of Austrosaurus mckillopi (QM F2316, holotype) in

2280 millimetres.

Measurements Left dorsal ribs I II III IV V Preserved proximodistal length 781 1347 1379 1484 1702

2281 Table 2. Measurements of the dorsal ribs of Austrosaurus mckillopi (QM F2316, holotype) in

2282 millimetres.

Measurement QM F6142 Centrum Preserved anteroposterior length 279 Posterior (cotyle) Maximum height 199 Maximum width 275 Pneumatic fossa Maximum length 195 Maximum height 74 Postzygapophyses Combined transverse width 383

102

2283 Table 3. Measurements of the cervical vertebra of the “Hughenden sauropod” (QM F6142) in

2284 millimetres.

103

2285 Appendix

2286 List of specimens catalogued under QM F2316, the Austrosaurus mckillopi hypodigm. For

2287 the purposes of continuity, rather than convenience, the holotypic blocks initially described

2288 by Longman (1933) are referred to by their original letters.

Specimen Description Previous references Notes A Two incomplete yet articulated “Specimen A” in Sequentially dorsal centra representing the Longman (1933, fig. posterior to posterior portion of dorsal 3 and pls. XV and specimen B and vertebra II (a1) and the anterior XVI); “Block A” in anterior to specimen portion of dorsal vertebra III Molnar (2001b, fig. H. Large block (a2). One fragment, which 1); “Longman’s (~255 mm long) preserves portions of the neural specimen A” in with a large arches of both dorsal vertebrae II Molnar & Salisbury detachable fragment and III (a3) keys into the dorsal (2005, fig. 20.1A). comprising mostly surface of the combined a1+a2 matrix. block and includes a rib fragment. B Two incomplete yet articulated “Specimen B” in Sequentially dorsal centra representing the Longman (1933, pl. posterior to posterior portion of dorsal XVII). specimen C and vertebra I (b1) and the anterior anterior to specimen portion of dorsal vertebra II (b2). A. Large block A portion of matrix which (~300 mm long × preserves a small rib fragment ~280 mm high), with (b3) keys into the left lateral two smaller pneumatic fossa of b2. A fourth detachable fragment, comprising a neural fragments. canal cast (b4), keys into the dorsal margin of the intersection between b1 and b2. C Two incomplete yet articulated “Specimen C” in Sequentially anterior presacral centra representing the Longman (1933). to specimen B. posterior portion of the Large block (280 posteriormost cervical vertebra mm wide × 204 mm (c1) and the anterior portion of high) dorsal vertebra I (c2). D Mid-section of dorsal vertebra IV Molnar & Salisbury Sequentially (2005, fig. 20.1B–C); posterior to note that “Longman’s specimen E and (1933) Specimen C” anterior to specimen in Molnar (2011a, fig. F. Large block.

104

1H) is not Longman’s Specimen C. E Two incomplete yet articulated Molnar & Salisbury Sequentially dorsal centra representing the (2005, fig. 20.7). posterior to posterior portion of dorsal specimen H and vertebra III (e1) and the anterior anterior to specimen portion of dorsal vertebra IV D. Large block. (e2). F Two incomplete yet articulated Sequentially dorsal centra representing the posterior to posteriormost portion of dorsal specimen D. Large vertebra IV (f1) and the anterior block. portion of dorsal vertebra V (f2). G A fragment of mostly internal Keys into specimens tissue around presumed neural N and W. Large canal infill. block. H Thin centrum portion from dorsal Sequentially vertebra III. posterior to specimen A and anterior to specimen E. Small fragment. I Indeterminate fragment of Small fragment (145 internal tissue, containing mm × 102 mm). possible laminae. J Right posterior section of a dorsal Small fragment (119 vertebra. mm × 88 mm). K Neural canal cast with Small fragment (70 indeterminate internal tissue mm × 50 mm). L Rib fragment Small fragment. M Vertebral transverse process Small fragment (90 mm × ~110 mm–140 mm). N Indeterminate vertebral fragment Keys into specimen G. Small fragment (135 mm × 200 mm). O Vertebral centrum fragment Small fragment (98 mm × 150 mm long. P Vertebral inter-zygapophyseal Small fragment (120 components of adjoining mm × 107 mm × 100 vertebrae. mm). Q Vertebral inter-centrum Small fragment (109 components of adjoining mm × 68 mm). vertebrae. R Vertebral inter-centrum Small fragment (102 components of adjoining mm × 63 mm). vertebrae.

105

S Matrix with two indeterminate Small fragment (64 fragments. mm × 22 mm; 101 mm × 51 mm). T Matrix with two indeterminate Small fragment (83 fragments. mm × 28 mm; 130 mm × 18 mm). U Indeterminate presacral vertebral Small fragment (120 fragment. mm × 112 mm). V Indeterminate fragment, possibly Small fragment. of a rib. W Indeterminate fragment, mostly Keys into specimen comprising internal issue. G. Small fragment (133 mm × 150 mm). X Fragment of a neural arch, Small fragment. containing a section of neural canal cast. Y Condylar fragment of a centrum. Molnar (2011a, fig. Small fragment (181 1I) mm × 106 mm). 2289